RFC 4728






Network Working Group                                         D. Johnson
Request for Comments: 4728                               Rice University
Category: Experimental                                             Y. Hu
                                                                    UIUC
                                                                D. Maltz
                                                      Microsoft Research
                                                           February 2007


               The Dynamic Source Routing Protocol (DSR)
                  for Mobile Ad Hoc Networks for IPv4

Status of This Memo



   This memo defines an Experimental Protocol for the Internet
   community.  It does not specify an Internet standard of any kind.
   Discussion and suggestions for improvement are requested.
   Distribution of this memo is unlimited.

Copyright Notice



   Copyright (C) The IETF Trust (2007).

Abstract



   The Dynamic Source Routing protocol (DSR) is a simple and efficient
   routing protocol designed specifically for use in multi-hop wireless
   ad hoc networks of mobile nodes.  DSR allows the network to be
   completely self-organizing and self-configuring, without the need for
   any existing network infrastructure or administration.  The protocol
   is composed of the two main mechanisms of "Route Discovery" and
   "Route Maintenance", which work together to allow nodes to discover
   and maintain routes to arbitrary destinations in the ad hoc network.
   All aspects of the protocol operate entirely on demand, allowing the
   routing packet overhead of DSR to scale automatically to only what is
   needed to react to changes in the routes currently in use.  The
   protocol allows multiple routes to any destination and allows each
   sender to select and control the routes used in routing its packets,
   for example, for use in load balancing or for increased robustness.
   Other advantages of the DSR protocol include easily guaranteed loop-
   free routing, operation in networks containing unidirectional links,
   use of only "soft state" in routing, and very rapid recovery when
   routes in the network change.  The DSR protocol is designed mainly
   for mobile ad hoc networks of up to about two hundred nodes and is
   designed to work well even with very high rates of mobility.  This
   document specifies the operation of the DSR protocol for routing
   unicast IPv4 packets.




Johnson, et al.               Experimental                      [Page 1]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Table of Contents



   1. Introduction ....................................................5
   2. Assumptions .....................................................7
   3. DSR Protocol Overview ...........................................9
      3.1. Basic DSR Route Discovery .................................10
      3.2. Basic DSR Route Maintenance ...............................12
      3.3. Additional Route Discovery Features .......................14
           3.3.1. Caching Overheard Routing Information ..............14
           3.3.2. Replying to Route Requests Using Cached Routes .....15
           3.3.3. Route Request Hop Limits ...........................16
      3.4. Additional Route Maintenance Features .....................17
           3.4.1. Packet Salvaging ...................................17
           3.4.2. Queued Packets Destined over a Broken Link .........18
           3.4.3. Automatic Route Shortening .........................19
           3.4.4. Increased Spreading of Route Error Messages ........20
      3.5. Optional DSR Flow State Extension .........................20
           3.5.1. Flow Establishment .................................21
           3.5.2. Receiving and Forwarding Establishment Packets .....22
           3.5.3. Sending Packets along Established Flows ............22
           3.5.4. Receiving and Forwarding Packets Sent along
                  Established Flows ..................................23
           3.5.5. Processing Route Errors ............................24
           3.5.6. Interaction with Automatic Route Shortening ........24
           3.5.7. Loop Detection .....................................25
           3.5.8. Acknowledgement Destination ........................25
           3.5.9. Crash Recovery .....................................25
           3.5.10. Rate Limiting .....................................25
           3.5.11. Interaction with Packet Salvaging .................26
   4. Conceptual Data Structures .....................................26
      4.1. Route Cache ...............................................26
      4.2. Send Buffer ...............................................30
      4.3. Route Request Table .......................................30
      4.4. Gratuitous Route Reply Table ..............................31
      4.5. Network Interface Queue and Maintenance Buffer ............32
      4.6. Blacklist .................................................33
   5. Additional Conceptual Data Structures for Flow State
      Extension ......................................................34
      5.1. Flow Table ................................................34
      5.2. Automatic Route Shortening Table ..........................35
      5.3. Default Flow ID Table .....................................36
   6. DSR Options Header Format ......................................36
      6.1. Fixed Portion of DSR Options Header .......................37
      6.2. Route Request Option ......................................40
      6.3. Route Reply Option ........................................42






Johnson, et al.               Experimental                      [Page 2]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      6.4. Route Error Option ........................................44
           6.4.1. Node Unreachable Type-Specific Information .........46
           6.4.2. Flow State Not Supported Type-Specific
                  Information ........................................46
           6.4.3. Option Not Supported Type-Specific Information .....46
      6.5. Acknowledgement Request Option ............................46
      6.6. Acknowledgement Option ....................................47
      6.7. DSR Source Route Option ...................................48
      6.8. Pad1 Option ...............................................50
      6.9. PadN Option ...............................................50
   7. Additional Header Formats and Options for Flow State
      Extension ......................................................51
      7.1. DSR Flow State Header .....................................52
      7.2. New Options and Extensions in DSR Options Header ..........52
           7.2.1. Timeout Option .....................................52
           7.2.2. Destination and Flow ID Option .....................53
      7.3. New Error Types for Route Error Option ....................54
           7.3.1. Unknown Flow Type-Specific Information .............54
           7.3.2. Default Flow Unknown Type-Specific Information .....55
      7.4. New Acknowledgement Request Option Extension ..............55
           7.4.1. Previous Hop Address Extension .....................55
   8. Detailed Operation .............................................56
      8.1. General Packet Processing .................................56
           8.1.1. Originating a Packet ...............................56
           8.1.2. Adding a DSR Options Header to a Packet ............57
           8.1.3. Adding a DSR Source Route Option to a Packet .......57
           8.1.4. Processing a Received Packet .......................58
           8.1.5. Processing a Received DSR Source Route Option ......60
           8.1.6. Handling an Unknown DSR Option .....................63
      8.2. Route Discovery Processing ................................64
           8.2.1. Originating a Route Request ........................65
           8.2.2. Processing a Received Route Request Option .........66
           8.2.3. Generating a Route Reply Using the Route Cache .....68
           8.2.4. Originating a Route Reply ..........................71
           8.2.5. Preventing Route Reply Storms ......................72
           8.2.6. Processing a Received Route Reply Option ...........74
      8.3. Route Maintenance Processing ..............................74
           8.3.1. Using Link-Layer Acknowledgements ..................75
           8.3.2. Using Passive Acknowledgements .....................76
           8.3.3. Using Network-Layer Acknowledgements ...............77
           8.3.4. Originating a Route Error ..........................80
           8.3.5. Processing a Received Route Error Option ...........81
           8.3.6. Salvaging a Packet .................................82
      8.4. Multiple Network Interface Support ........................84
      8.5. IP Fragmentation and Reassembly ...........................84
      8.6. Flow State Processing .....................................85
           8.6.1. Originating a Packet ...............................85
           8.6.2. Inserting a DSR Flow State Header ..................88



Johnson, et al.               Experimental                      [Page 3]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


           8.6.3. Receiving a Packet .................................88
           8.6.4. Forwarding a Packet Using Flow IDs .................93
           8.6.5. Promiscuously Receiving a Packet ...................93
           8.6.6. Operation Where the Layer below DSR
                  Decreases the IP TTL ...............................94
           8.6.7. Salvage Interactions with DSR ......................94
   9. Protocol Constants and Configuration Variables .................95
   10. IANA Considerations ...........................................96
   11. Security Considerations .......................................96
   Appendix A. Link-MaxLife Cache Description ........................97
   Appendix B. Location of DSR in the ISO Network Reference Model ....99
   Appendix C. Implementation and Evaluation Status .................100
   Acknowledgements .................................................101
   Normative References .............................................102
   Informative References ...........................................102




































Johnson, et al.               Experimental                      [Page 4]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


1.  Introduction



   The Dynamic Source Routing protocol (DSR) [JOHNSON94, JOHNSON96a] is
   a simple and efficient routing protocol designed specifically for use
   in multi-hop wireless ad hoc networks of mobile nodes.  Using DSR,
   the network is completely self-organizing and self-configuring,
   requiring no existing network infrastructure or administration.
   Network nodes cooperate to forward packets for each other to allow
   communication over multiple "hops" between nodes not directly within
   wireless transmission range of one another.  As nodes in the network
   move about or join or leave the network, and as wireless transmission
   conditions such as sources of interference change, all routing is
   automatically determined and maintained by the DSR routing protocol.
   Since the number or sequence of intermediate hops needed to reach any
   destination may change at any time, the resulting network topology
   may be quite rich and rapidly changing.

   In designing DSR, we sought to create a routing protocol that had
   very low overhead yet was able to react very quickly to changes in
   the network.  The DSR protocol provides highly reactive service in
   order to help ensure successful delivery of data packets in spite of
   node movement or other changes in network conditions.

   The DSR protocol is composed of two main mechanisms that work
   together to allow the discovery and maintenance of source routes in
   the ad hoc network:

   -  Route Discovery is the mechanism by which a node S wishing to send
      a packet to a destination node D obtains a source route to D.
      Route Discovery is used only when S attempts to send a packet to D
      and does not already know a route to D.

   -  Route Maintenance is the mechanism by which node S is able to
      detect, while using a source route to D, if the network topology
      has changed such that it can no longer use its route to D because
      a link along the route no longer works.  When Route Maintenance
      indicates a source route is broken, S can attempt to use any other
      route it happens to know to D, or it can invoke Route Discovery
      again to find a new route for subsequent packets to D.  Route
      Maintenance for this route is used only when S is actually sending
      packets to D.

   In DSR, Route Discovery and Route Maintenance each operate entirely
   "on demand".  In particular, unlike other protocols, DSR requires no
   periodic packets of any kind at any layer within the network.  For
   example, DSR does not use any periodic routing advertisement, link
   status sensing, or neighbor detection packets and does not rely on
   these functions from any underlying protocols in the network.  This



Johnson, et al.               Experimental                      [Page 5]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   entirely on-demand behavior and lack of periodic activity allows the
   number of overhead packets caused by DSR to scale all the way down to
   zero, when all nodes are approximately stationary with respect to
   each other and all routes needed for current communication have
   already been discovered.  As nodes begin to move more or as
   communication patterns change, the routing packet overhead of DSR
   automatically scales to only what is needed to track the routes
   currently in use.  Network topology changes not affecting routes
   currently in use are ignored and do not cause reaction from the
   protocol.

   All state maintained by DSR is "soft state" [CLARK88], in that the
   loss of any state will not interfere with the correct operation of
   the protocol; all state is discovered as needed and can easily and
   quickly be rediscovered if needed after a failure without significant
   impact on the protocol.  This use of only soft state allows the
   routing protocol to be very robust to problems such as dropped or
   delayed routing packets or node failures.  In particular, a node in
   DSR that fails and reboots can easily rejoin the network immediately
   after rebooting; if the failed node was involved in forwarding
   packets for other nodes as an intermediate hop along one or more
   routes, it can also resume this forwarding quickly after rebooting,
   with no or minimal interruption to the routing protocol.

   In response to a single Route Discovery (as well as through routing
   information from other packets overheard), a node may learn and cache
   multiple routes to any destination.  This support for multiple routes
   allows the reaction to routing changes to be much more rapid, since a
   node with multiple routes to a destination can try another cached
   route if the one it has been using should fail.  This caching of
   multiple routes also avoids the overhead of needing to perform a new
   Route Discovery each time a route in use breaks.  The sender of a
   packet selects and controls the route used for its own packets,
   which, together with support for multiple routes, also allows
   features such as load balancing to be defined.  In addition, all
   routes used are easily guaranteed to be loop-free, since the sender
   can avoid duplicate hops in the routes selected.

   The operation of both Route Discovery and Route Maintenance in DSR
   are designed to allow unidirectional links and asymmetric routes to
   be supported.  In particular, as noted in Section 2, in wireless
   networks, it is possible that a link between two nodes may not work
   equally well in both directions, due to differing transmit power
   levels or sources of interference.

   It is possible to interface a DSR network with other networks,
   external to this DSR network.  Such external networks may, for
   example, be the Internet or may be other ad hoc networks routed with



Johnson, et al.               Experimental                      [Page 6]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   a routing protocol other than DSR.  Such external networks may also
   be other DSR networks that are treated as external networks in order
   to improve scalability.  The complete handling of such external
   networks is beyond the scope of this document.  However, this
   document specifies a minimal set of requirements and features
   necessary to allow nodes only implementing this specification to
   interoperate correctly with nodes implementing interfaces to such
   external networks.

   This document specifies the operation of the DSR protocol for routing
   unicast IPv4 packets in multi-hop wireless ad hoc networks.
   Advanced, optional features, such as Quality of Service (QoS) support
   and efficient multicast routing, and operation of DSR with IPv6
   [RFC2460], will be covered in other documents.  The specification of
   DSR in this document provides a compatible base on which such
   features can be added, either independently or by integration with
   the DSR operation specified here.  As described in Appendix C, the
   design of DSR has been extensively studied through detailed
   simulations and testbed implementation and demonstration; this
   document encourages additional implementation and experimentation
   with the protocol.

   The keywords "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in RFC 2119 [RFC2119].

2.  Assumptions



   As described here, the DSR protocol is designed mainly for mobile ad
   hoc networks of up to about two hundred nodes and is designed to work
   well even with very high rates of mobility.  Other protocol features
   and enhancements that may allow DSR to scale to larger networks are
   outside the scope of this document.

   We assume in this document that all nodes wishing to communicate with
   other nodes within the ad hoc network are willing to participate
   fully in the protocols of the network.  In particular, each node
   participating in the ad hoc network SHOULD also be willing to forward
   packets for other nodes in the network.

   The diameter of an ad hoc network is the minimum number of hops
   necessary for a packet to reach from any node located at one extreme
   edge of the ad hoc network to another node located at the opposite
   extreme.  We assume that this diameter will often be small (e.g.,
   perhaps 5 or 10 hops), but it may often be greater than 1.






Johnson, et al.               Experimental                      [Page 7]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   Packets may be lost or corrupted in transmission on the wireless
   network.  We assume that a node receiving a corrupted packet can
   detect the error, such as through a standard link-layer checksum or
   Cyclic Redundancy Check (CRC), and discard the packet.

   Nodes within the ad hoc network MAY move at any time without notice
   and MAY even move continuously, but we assume that the speed with
   which nodes move is moderate with respect to the packet transmission
   latency and wireless transmission range of the particular underlying
   network hardware in use.  In particular, DSR can support very rapid
   rates of arbitrary node mobility, but we assume that nodes do not
   continuously move so rapidly as to make the flooding of every
   individual data packet the only possible routing protocol.

   A common feature of many network interfaces, including most current
   LAN hardware for broadcast media such as wireless, is the ability to
   operate the network interface in "promiscuous" receive mode.  This
   mode causes the hardware to deliver every received packet to the
   network driver software without filtering based on link-layer
   destination address.  Although we do not require this facility, some
   of our optimizations can take advantage of its availability.  Use of
   promiscuous mode does increase the software overhead on the CPU, but
   we believe that wireless network speeds and capacity are more the
   inherent limiting factors to performance in current and future
   systems; we also believe that portions of the protocol are suitable
   for implementation directly within a programmable network interface
   unit to avoid this overhead on the CPU [JOHNSON96a].  Use of
   promiscuous mode may also increase the power consumption of the
   network interface hardware, depending on the design of the receiver
   hardware, and in such cases, DSR can easily be used without the
   optimizations that depend on promiscuous receive mode or can be
   programmed to only periodically switch the interface into promiscuous
   mode.  Use of promiscuous receive mode is entirely optional.

   Wireless communication ability between any pair of nodes may at times
   not work equally well in both directions, due, for example, to
   transmit power levels or sources of interference around the two nodes
   [BANTZ94, LAUER95].  That is, wireless communications between each
   pair of nodes will in many cases be able to operate bidirectionally,
   but at times the wireless link between two nodes may be only
   unidirectional, allowing one node to successfully send packets to the
   other while no communication is possible in the reverse direction.
   Some Medium Access Control (MAC) protocols, however, such as MACA
   [KARN90], MACAW [BHARGHAVAN94], or IEEE 802.11 [IEEE80211], limit
   unicast data packet transmission to bidirectional links, due to the
   required bidirectional exchange of request to send (RTS) and clear to
   send (CTS) packets in these protocols and to the link-layer
   acknowledgement feature in IEEE 802.11.  When used on top of MAC



Johnson, et al.               Experimental                      [Page 8]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   protocols such as these, DSR can take advantage of additional
   optimizations, such as the ability to reverse a source route to
   obtain a route back to the origin of the original route.

   The IP address used by a node using the DSR protocol MAY be assigned
   by any mechanism (e.g., static assignment or use of Dynamic Host
   Configuration Protocol (DHCP) for dynamic assignment [RFC2131]),
   although the method of such assignment is outside the scope of this
   specification.

   A routing protocol such as DSR chooses a next-hop for each packet and
   provides the IP address of that next-hop.  When the packet is
   transmitted, however, the lower-layer protocol often has a separate,
   MAC-layer address for the next-hop node.  DSR uses the Address
   Resolution Protocol (ARP) [RFC826] to translate from next-hop IP
   addresses to next-hop MAC addresses.  In addition, a node MAY add an
   entry to its ARP cache based on any received packet, when the IP
   address and MAC address of the transmitting node are available in the
   packet; for example, the IP address of the transmitting node is
   present in a Route Request option (in the Address list being
   accumulated) and any packets containing a source route.  Adding
   entries to the ARP cache in this way avoids the overhead of ARP in
   most cases.

3.  DSR Protocol Overview



   This section provides an overview of the operation of the DSR
   protocol.  The basic version of DSR uses explicit "source routing",
   in which each data packet sent carries in its header the complete,
   ordered list of nodes through which the packet will pass.  This use
   of explicit source routing allows the sender to select and control
   the routes used for its own packets, supports the use of multiple
   routes to any destination (for example, for load balancing), and
   allows a simple guarantee that the routes used are loop-free.  By
   including this source route in the header of each data packet, other
   nodes forwarding or overhearing any of these packets can also easily
   cache this routing information for future use.  Section 3.1 describes
   this basic operation of Route Discovery, Section 3.2 describes basic
   Route Maintenance, and Sections 3.3 and 3.4 describe additional
   features of these two parts of DSR's operation.  Section 3.5 then
   describes an optional, compatible extension to DSR, known as "flow
   state", that allows the routing of most packets without an explicit
   source route header in the packet, while the fundamental properties
   of DSR's operation are preserved.







Johnson, et al.               Experimental                      [Page 9]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


3.1.  Basic DSR Route Discovery



   When some source node originates a new packet addressed to some
   destination node, the source node places in the header of the packet
   a "source route" giving the sequence of hops that the packet is to
   follow on its way to the destination.  Normally, the sender will
   obtain a suitable source route by searching its "Route Cache" of
   routes previously learned; if no route is found in its cache, it will
   initiate the Route Discovery protocol to dynamically find a new route
   to this destination node.  In this case, we call the source node the
   "initiator" and the destination node the "target" of the Route
   Discovery.

   For example, suppose a node A is attempting to discover a route to
   node E.  The Route Discovery initiated by node A in this example
   would proceed as follows:

            ^    "A"    ^   "A,B"   ^  "A,B,C"  ^ "A,B,C,D"
            |   id=2    |   id=2    |   id=2    |   id=2
         +-----+     +-----+     +-----+     +-----+     +-----+
         |  A  |---->|  B  |---->|  C  |---->|  D  |---->|  E  |
         +-----+     +-----+     +-----+     +-----+     +-----+
            |           |           |           |
            v           v           v           v

   To initiate the Route Discovery, node A transmits a "Route Request"
   as a single local broadcast packet, which is received by
   (approximately) all nodes currently within wireless transmission
   range of A, including node B in this example.  Each Route Request
   identifies the initiator and target of the Route Discovery, and also
   contains a unique request identification (2, in this example),
   determined by the initiator of the Request.  Each Route Request also
   contains a record listing the address of each intermediate node
   through which this particular copy of the Route Request has been
   forwarded.  This route record is initialized to an empty list by the
   initiator of the Route Discovery.  In this example, the route record
   initially lists only node A.

   When another node receives this Route Request (such as node B in this
   example), if it is the target of the Route Discovery, it returns a
   "Route Reply" to the initiator of the Route Discovery, giving a copy
   of the accumulated route record from the Route Request; when the
   initiator receives this Route Reply, it caches this route in its
   Route Cache for use in sending subsequent packets to this
   destination.






Johnson, et al.               Experimental                     [Page 10]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   Otherwise, if this node receiving the Route Request has recently seen
   another Route Request message from this initiator bearing this same
   request identification and target address, or if this node's own
   address is already listed in the route record in the Route Request,
   this node discards the Request.  (A node considers a Request recently
   seen if it still has information about that Request in its Route
   Request Table, which is described in Section 4.3.)  Otherwise, this
   node appends its own address to the route record in the Route Request
   and propagates it by transmitting it as a local broadcast packet
   (with the same request identification).  In this example, node B
   broadcast the Route Request, which is received by node C; nodes C and
   D each also, in turn, broadcast the Request, resulting in receipt of
   a copy of the Request by node E.

   In returning the Route Reply to the initiator of the Route Discovery,
   such as in this example, node E replying back to node A, node E will
   typically examine its own Route Cache for a route back to A and, if
   one is found, will use it for the source route for delivery of the
   packet containing the Route Reply.  Otherwise, E SHOULD perform its
   own Route Discovery for target node A, but to avoid possible infinite
   recursion of Route Discoveries, it MUST in this case piggyback this
   Route Reply on the packet containing its own Route Request for A.  It
   is also possible to piggyback other small data packets, such as a TCP
   SYN packet [RFC793], on a Route Request using this same mechanism.

   Node E could instead simply reverse the sequence of hops in the route
   record that it is trying to send in the Route Reply and use this as
   the source route on the packet carrying the Route Reply itself.  For
   MAC protocols, such as IEEE 802.11, that require a bidirectional
   frame exchange for unicast packets as part of the MAC protocol
   [IEEE80211], the discovered source route MUST be reversed in this way
   to return the Route Reply, since this route reversal tests the
   discovered route to ensure that it is bidirectional before the Route
   Discovery initiator begins using the route.  This route reversal also
   avoids the overhead of a possible second Route Discovery.

   When initiating a Route Discovery, the sending node saves a copy of
   the original packet (that triggered the discovery) in a local buffer
   called the "Send Buffer".  The Send Buffer contains a copy of each
   packet that cannot be transmitted by this node because it does not
   yet have a source route to the packet's destination.  Each packet in
   the Send Buffer is logically associated with the time that it was
   placed into the Send Buffer and is discarded after residing in the
   Send Buffer for some timeout period SendBufferTimeout; if necessary
   for preventing the Send Buffer from overflowing, a FIFO or other
   replacement strategy MAY also be used to evict packets even before
   they expire.




Johnson, et al.               Experimental                     [Page 11]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   While a packet remains in the Send Buffer, the node SHOULD
   occasionally initiate a new Route Discovery for the packet's
   destination address.  However, the node MUST limit the rate at which
   such new Route Discoveries for the same address are initiated (as
   described in Section 4.3), since it is possible that the destination
   node is not currently reachable.  In particular, due to the limited
   wireless transmission range and the movement of the nodes in the
   network, the network may at times become partitioned, meaning that
   there is currently no sequence of nodes through which a packet could
   be forwarded to reach the destination.  Depending on the movement
   pattern and the density of nodes in the network, such network
   partitions may be rare or common.

   If a new Route Discovery was initiated for each packet sent by a node
   in such a partitioned network, a large number of unproductive Route
   Request packets would be propagated throughout the subset of the ad
   hoc network reachable from this node.  In order to reduce the
   overhead from such Route Discoveries, a node SHOULD use an
   exponential back-off algorithm to limit the rate at which it
   initiates new Route Discoveries for the same target, doubling the
   timeout between each successive discovery initiated for the same
   target.  If the node attempts to send additional data packets to this
   same destination node more frequently than this limit, the subsequent
   packets SHOULD be buffered in the Send Buffer until a Route Reply is
   received giving a route to this destination, but the node MUST NOT
   initiate a new Route Discovery until the minimum allowable interval
   between new Route Discoveries for this target has been reached.  This
   limitation on the maximum rate of Route Discoveries for the same
   target is similar to the mechanism required by Internet nodes to
   limit the rate at which ARP Requests are sent for any single target
   IP address [RFC1122].

3.2.  Basic DSR Route Maintenance



   When originating or forwarding a packet using a source route, each
   node transmitting the packet is responsible for confirming that data
   can flow over the link from that node to the next hop.  For example,
   in the situation shown below, node A has originated a packet for node
   E using a source route through intermediate nodes B, C, and D:

         +-----+     +-----+     +-----+     +-----+     +-----+
         |  A  |---->|  B  |---->|  C  |-->? |  D  |     |  E  |
         +-----+     +-----+     +-----+     +-----+     +-----+

   In this case, node A is responsible for the link from A to B, node B
   is responsible for the link from B to C, node C is responsible for
   the link from C to D, and node D is responsible for the link from D
   to E.



Johnson, et al.               Experimental                     [Page 12]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   An acknowledgement can provide confirmation that a link is capable of
   carrying data, and in wireless networks, acknowledgements are often
   provided at no cost, either as an existing standard part of the MAC
   protocol in use (such as the link-layer acknowledgement frame defined
   by IEEE 802.11 [IEEE80211]), or by a "passive acknowledgement"
   [JUBIN87] (in which, for example, B confirms receipt at C by
   overhearing C transmit the packet when forwarding it on to D).

   If a built-in acknowledgement mechanism is not available, the node
   transmitting the packet can explicitly request that a DSR-specific
   software acknowledgement be returned by the next node along the
   route; this software acknowledgement will normally be transmitted
   directly to the sending node, but if the link between these two nodes
   is unidirectional (Section 4.6), this software acknowledgement could
   travel over a different, multi-hop path.

   After an acknowledgement has been received from some neighbor, a node
   MAY choose not to require acknowledgements from that neighbor for a
   brief period of time, unless the network interface connecting a node
   to that neighbor always receives an acknowledgement in response to
   unicast traffic.

   When a software acknowledgement is used, the acknowledgement request
   SHOULD be retransmitted up to a maximum number of times.  A
   retransmission of the acknowledgement request can be sent as a
   separate packet, piggybacked on a retransmission of the original data
   packet, or piggybacked on any packet with the same next-hop
   destination that does not also contain a software acknowledgement.

   After the acknowledgement request has been retransmitted the maximum
   number of times, if no acknowledgement has been received, then the
   sender treats the link to this next-hop destination as currently
   "broken".  It SHOULD remove this link from its Route Cache and SHOULD
   return a "Route Error" to each node that has sent a packet routed
   over that link since an acknowledgement was last received.  For
   example, in the situation shown above, if C does not receive an
   acknowledgement from D after some number of requests, it would return
   a Route Error to A, as well as any other node that may have used the
   link from C to D since C last received an acknowledgement from D.
   Node A then removes this broken link from its cache; any
   retransmission of the original packet can be performed by upper layer
   protocols such as TCP, if necessary.  For sending such a
   retransmission or other packets to this same destination E, if A has
   in its Route Cache another route to E (for example, from additional
   Route Replies from its earlier Route Discovery, or from having
   overheard sufficient routing information from other packets), it can





Johnson, et al.               Experimental                     [Page 13]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   send the packet using the new route immediately.  Otherwise, it
   SHOULD perform a new Route Discovery for this target (subject to the
   back-off described in Section 3.1).

3.3.  Additional Route Discovery Features



3.3.1.  Caching Overheard Routing Information



   A node forwarding or otherwise overhearing any packet SHOULD add all
   usable routing information from that packet to its own Route Cache.
   The usefulness of routing information in a packet depends on the
   directionality characteristics of the physical medium (Section 2), as
   well as on the MAC protocol being used.  Specifically, three distinct
   cases are possible:

   -  Links in the network frequently are capable of operating only
      unidirectionally (not bidirectionally), and the MAC protocol in
      use in the network is capable of transmitting unicast packets over
      unidirectional links.

   -  Links in the network occasionally are capable of operating only
      unidirectionally (not bidirectionally), but this unidirectional
      restriction on any link is not persistent; almost all links are
      physically bidirectional, and the MAC protocol in use in the
      network is capable of transmitting unicast packets over
      unidirectional links.

   -  The MAC protocol in use in the network is not capable of
      transmitting unicast packets over unidirectional links; only
      bidirectional links can be used by the MAC protocol for
      transmitting unicast packets.  For example, the IEEE 802.11
      Distributed Coordination Function (DCF) MAC protocol [IEEE80211]
      is capable of transmitting a unicast packet only over a
      bidirectional link, since the MAC protocol requires the return of
      a link-level acknowledgement packet from the receiver and also
      optionally requires the bidirectional exchange of an RTS and CTS
      packet between the transmitter and receiver nodes.

   In the first case above, for example, the source route used in a data
   packet, the accumulated route record in a Route Request, or the route
   being returned in a Route Reply SHOULD all be cached by any node in
   the "forward" direction.  Any node SHOULD cache this information from
   any such packet received, whether the packet was addressed to this
   node, sent to a broadcast (or multicast) MAC address, or overheard
   while the node's network interface is in promiscuous mode.  However,
   the "reverse" direction of the links identified in such packet
   headers SHOULD NOT be cached.




Johnson, et al.               Experimental                     [Page 14]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   For example, in the situation shown below, node A is using a source
   route to communicate with node E:

      +-----+     +-----+     +-----+     +-----+     +-----+
      |  A  |---->|  B  |---->|  C  |---->|  D  |---->|  E  |
      +-----+     +-----+     +-----+     +-----+     +-----+

   As node C forwards a data packet along the route from A to E, it
   SHOULD add to its cache the presence of the "forward" direction links
   that it learns from the headers of these packets, from itself to D
   and from D to E.  Node C SHOULD NOT, in this case, cache the
   "reverse" direction of the links identified in these packet headers,
   from itself back to B and from B to A, since these links might be
   unidirectional.

   In the second case above, in which links may occasionally operate
   unidirectionally, the links described above SHOULD be cached in both
   directions.  Furthermore, in this case, if node X overhears (e.g.,
   through promiscuous mode) a packet transmitted by node C that is
   using a source route from node A to E, node X SHOULD cache all of
   these links as well, also including the link from C to X over which
   it overheard the packet.

   In the final case, in which the MAC protocol requires physical
   bidirectionality for unicast operation, links from a source route
   SHOULD be cached in both directions, except when the packet also
   contains a Route Reply, in which case only the links already
   traversed in this source route SHOULD be cached.  However, the links
   not yet traversed in this route SHOULD NOT be cached.

3.3.2.  Replying to Route Requests Using Cached Routes



   A node receiving a Route Request for which it is not the target
   searches its own Route Cache for a route to the target of the
   Request.  If it is found, the node generally returns a Route Reply to
   the initiator itself rather than forward the Route Request.  In the
   Route Reply, this node sets the route record to list the sequence of
   hops over which this copy of the Route Request was forwarded to it,
   concatenated with the source route to this target obtained from its
   own Route Cache.

   However, before transmitting a Route Reply packet that was generated
   using information from its Route Cache in this way, a node MUST
   verify that the resulting route being returned in the Route Reply,
   after this concatenation, contains no duplicate nodes listed in the
   route record.  For example, the figure below illustrates a case in
   which a Route Request for target E has been received by node F, and
   node F already has in its Route Cache a route from itself to E:



Johnson, et al.               Experimental                     [Page 15]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


         +-----+     +-----+                 +-----+     +-----+
         |  A  |---->|  B  |-               >|  D  |---->|  E  |
         +-----+     +-----+ \             / +-----+     +-----+
                              \           /
                               \ +-----+ /
                                >|  C  |-
                                 +-----+
                                   | ^
                                   v |
           Route Request         +-----+
           Route: A - B - C - F  |  F  |  Cache: C - D - E
                                 +-----+

   The concatenation of the accumulated route record from the Route
   Request and the cached route from F's Route Cache would include a
   duplicate node in passing from C to F and back to C.

   Node F in this case could attempt to edit the route to eliminate the
   duplication, resulting in a route from A to B to C to D and on to E,
   but in this case, node F would not be on the route that it returned
   in its own Route Reply.  DSR Route Discovery prohibits node F from
   returning such a Route Reply from its cache; this prohibition
   increases the probability that the resulting route is valid, since
   node F in this case should have received a Route Error if the route
   had previously stopped working.  Furthermore, this prohibition means
   that a future Route Error traversing the route is very likely to pass
   through any node that sent the Route Reply for the route (including
   node F), which helps to ensure that stale data is removed from caches
   (such as at F) in a timely manner; otherwise, the next Route
   Discovery initiated by A might also be contaminated by a Route Reply
   from F containing the same stale route.  If, due to this restriction
   on returning a Route Reply based on information from its Route Cache,
   node F does not return such a Route Reply, it propagates the Route
   Request normally.

3.3.3.  Route Request Hop Limits



   Each Route Request message contains a "hop limit" that may be used to
   limit the number of intermediate nodes allowed to forward that copy
   of the Route Request.  This hop limit is implemented using the Time-
   to-Live (TTL) field in the IP header of the packet carrying the Route
   Request.  As the Request is forwarded, this limit is decremented, and
   the Request packet is discarded if the limit reaches zero before
   finding the target.  This Route Request hop limit can be used to
   implement a variety of algorithms for controlling the spread of a
   Route Request during a Route Discovery attempt.





Johnson, et al.               Experimental                     [Page 16]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   For example, a node MAY use this hop limit to implement a "non-
   propagating" Route Request as an initial phase of a Route Discovery.
   A node using this technique sends its first Route Request attempt for
   some target node using a hop limit of 1, such that any node receiving
   the initial transmission of the Route Request will not forward the
   Request to other nodes by re-broadcasting it.  This form of Route
   Request is called a "non-propagating" Route Request; it provides an
   inexpensive method for determining if the target is currently a
   neighbor of the initiator or if a neighbor node has a route to the
   target cached (effectively using the neighbors' Route Caches as an
   extension of the initiator's own Route Cache).  If no Route Reply is
   received after a short timeout, then the node sends a "propagating"
   Route Request for the target node (i.e., with hop limit as defined by
   the value of the DiscoveryHopLimit configuration variable).

   As another example, a node MAY use this hop limit to implement an
   "expanding ring" search for the target [JOHNSON96a].  A node using
   this technique sends an initial non-propagating Route Request as
   described above; if no Route Reply is received for it, the node
   originates another Route Request with a hop limit of 2.  For each
   Route Request originated, if no Route Reply is received for it, the
   node doubles the hop limit used on the previous attempt, to
   progressively explore for the target node without allowing the Route
   Request to propagate over the entire network.  However, this
   expanding ring search approach could increase the average latency of
   Route Discovery, since multiple Discovery attempts and timeouts may
   be needed before discovering a route to the target node.

3.4.  Additional Route Maintenance Features



3.4.1.  Packet Salvaging



   When an intermediate node forwarding a packet detects through Route
   Maintenance that the next hop along the route for that packet is
   broken, if the node has another route to the packet's destination in
   its Route Cache, the node SHOULD "salvage" the packet rather than
   discard it.  To salvage a packet, the node replaces the original
   source route on the packet with a route from its Route Cache.  The
   node then forwards the packet to the next node indicated along this
   source route.  For example, in the situation shown in the example of
   Section 3.2, if node C has another route cached to node E, it can
   salvage the packet by replacing the original route in the packet with
   this new route from its own Route Cache rather than discarding the
   packet.

   When salvaging a packet, a count is maintained in the packet of the
   number of times that it has been salvaged, to prevent a single packet
   from being salvaged endlessly.  Otherwise, since the TTL is



Johnson, et al.               Experimental                     [Page 17]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   decremented only once by each node, a single node could salvage a
   packet an unbounded number of times.  Even if we chose to require the
   TTL to be decremented on each salvage attempt, packet salvaging is an
   expensive operation, so it is desirable to bound the maximum number
   of times a packet can be salvaged independently of the maximum number
   of hops a packet can traverse.

   As described in Section 3.2, an intermediate node, such as in this
   case, that detects through Route Maintenance that the next hop along
   the route for a packet that it is forwarding is broken, the node also
   SHOULD return a Route Error to the original sender of the packet,
   identifying the link over which the packet could not be forwarded.
   If the node sends this Route Error, it SHOULD originate the Route
   Error before salvaging the packet.

3.4.2.  Queued Packets Destined over a Broken Link



   When an intermediate node forwarding a packet detects through Route
   Maintenance that the next-hop link along the route for that packet is
   broken, in addition to handling that packet as defined for Route
   Maintenance, the node SHOULD also handle in a similar way any pending
   packets that it has queued that are destined over this new broken
   link.  Specifically, the node SHOULD search its Network Interface
   Queue and Maintenance Buffer (Section 4.5) for packets for which the
   next-hop link is this new broken link.  For each such packet
   currently queued at this node, the node SHOULD process that packet as
   follows:

   -  Remove the packet from the node's Network Interface Queue and
      Maintenance Buffer.

   -  Originate a Route Error for this packet to the original sender of
      the packet, using the procedure described in Section 8.3.4, as if
      the node had already reached the maximum number of retransmission
      attempts for that packet for Route Maintenance.  However, in
      sending such Route Errors for queued packets in response to
      detection of a single, new broken link, the node SHOULD send no
      more than one Route Error to each original sender of any of these
      packets.

   -  If the node has another route to the packet's IP Destination
      Address in its Route Cache, the node SHOULD salvage the packet as
      described in Section 8.3.6.  Otherwise, the node SHOULD discard
      the packet.







Johnson, et al.               Experimental                     [Page 18]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


3.4.3.  Automatic Route Shortening



   Source routes in use MAY be automatically shortened if one or more
   intermediate nodes in the route become no longer necessary.  This
   mechanism of automatically shortening routes in use is somewhat
   similar to the use of passive acknowledgements [JUBIN87].  In
   particular, if a node is able to overhear a packet carrying a source
   route (e.g., by operating its network interface in promiscuous
   receive mode), then this node examines the unexpended portion of that
   source route.  If this node is not the intended next-hop destination
   for the packet but is named in the later unexpended portion of the
   packet's source route, then it can infer that the intermediate nodes
   before itself in the source route are no longer needed in the route.
   For example, the figure below illustrates an example in which node D
   has overheard a data packet being transmitted from B to C, for later
   forwarding to D and to E:

         +-----+     +-----+     +-----+     +-----+     +-----+
         |  A  |---->|  B  |---->|  C  |     |  D  |     |  E  |
         +-----+     +-----+     +-----+     +-----+     +-----+
                        \                       ^
                         \                     /
                          ---------------------

   In this case, this node (node D) SHOULD return a "gratuitous" Route
   Reply to the original sender of the packet (node A).  The Route Reply
   gives the shorter route as the concatenation of the portion of the
   original source route up through the node that transmitted the
   overheard packet (node B), plus the suffix of the original source
   route beginning with the node returning the gratuitous Route Reply
   (node D).  In this example, the route returned in the gratuitous
   Route Reply message sent from D to A gives the new route as the
   sequence of hops from A to B to D to E.

   When deciding whether to return a gratuitous Route Reply in this way,
   a node MAY factor in additional information beyond the fact that it
   was able to overhear the packet.  For example, the node MAY decide to
   return the gratuitous Route Reply only when the overheard packet is
   received with a signal strength or signal-to-noise ratio above some
   specific threshold.  In addition, each node maintains a Gratuitous
   Route Reply Table, as described in Section 4.4, to limit the rate at
   which it originates gratuitous Route Replies for the same returned
   route.








Johnson, et al.               Experimental                     [Page 19]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


3.4.4.  Increased Spreading of Route Error Messages



   When a source node receives a Route Error for a data packet that it
   originated, this source node propagates this Route Error to its
   neighbors by piggybacking it on its next Route Request.  In this way,
   stale information in the caches of nodes around this source node will
   not generate Route Replies that contain the same invalid link for
   which this source node received the Route Error.

   For example, in the situation shown in the example of Section 3.2,
   node A learns from the Route Error message from C that the link from
   C to D is currently broken.  It thus removes this link from its own
   Route Cache and initiates a new Route Discovery (if it has no other
   route to E in its Route Cache).  On the Route Request packet
   initiating this Route Discovery, node A piggybacks a copy of this
   Route Error, ensuring that the Route Error spreads well to other
   nodes, and guaranteeing that any Route Reply that it receives
   (including those from other node's Route Caches) in response to this
   Route Request does not contain a route that assumes the existence of
   this broken link.

3.5.  Optional DSR Flow State Extension



   This section describes an optional, compatible extension to the DSR
   protocol, known as "flow state", that allows the routing of most
   packets without an explicit source route header in the packet.  The
   DSR flow state extension further reduces the overhead of the protocol
   yet still preserves the fundamental properties of DSR's operation.
   Once a sending node has discovered a source route such as through
   DSR's Route Discovery mechanism, the flow state mechanism allows the
   sending node to establish hop-by-hop forwarding state within the
   network, based on this source route, to enable each node along the
   route to forward the packet to the next hop based on the node's own
   local knowledge of the flow along which this packet is being routed.
   Flow state is dynamically initialized by the first packet using a
   source route and is then able to route subsequent packets along the
   same flow without use of a source route header in the packet.  The
   state established at each hop along a flow is "soft state" and thus
   automatically expires when no longer needed and can be quickly
   recreated as necessary.  Extending DSR's basic operation based on an
   explicit source route in the header of each packet routed, the flow
   state extension operates as a form of "implicit source routing" by
   preserving DSR's basic operation but removing the explicit source
   route from packets.







Johnson, et al.               Experimental                     [Page 20]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


3.5.1.  Flow Establishment



   A source node sending packets to some destination node MAY use the
   DSR flow state extension described here to establish a route to that
   destination as a flow.  A "flow" is a route from the source to the
   destination represented by hop-by-hop forwarding state within the
   nodes along the route.  Each flow is uniquely identified by a
   combination of the source node address, the destination node address,
   and a flow identifier (flow ID) chosen by the source node.

   Each flow ID is a 16-bit unsigned integer.  Comparison between
   different flow IDs MUST be performed modulo 2**16.  For example,
   using an implementation in the C programming language, a flow ID
   value (a) is greater than another flow ID value (b) if
   ((short)((a) - (b)) > 0), if a C language "short" data type is
   implemented as a 16-bit signed integer.

   A DSR Flow State header in a packet identifies the flow ID to be
   followed in forwarding that packet.  From a given source to some
   destination, any number of different flows MAY exist and be in use,
   for example, following different sequences of hops to reach the
   destination.  One of these flows MAY be considered the "default" flow
   from that source to that destination.  If a node receives a packet
   with neither a DSR Options header specifying the route to be taken
   (with a Source Route option in the DSR Options header) nor a DSR Flow
   State header specifying the flow ID to be followed, it is forwarded
   along the default flow for the source and destination addresses
   specified in the packet's IP header.

   In establishing a new flow, the source node generates a nonzero
   16-bit flow ID greater than any unexpired flow IDs for this (source,
   destination) pair.  If the source wishes for this flow to become the
   default flow, the low bit of the flow ID MUST be set (the flow ID is
   an odd number); otherwise, the low bit MUST NOT be set (the flow ID
   is an even number).

   The source node establishing the new flow then transmits a packet
   containing a DSR Options header with a Source Route option.  To
   establish the flow, the source node also MUST include in the packet a
   DSR Flow State header, with the Flow ID field set to the chosen flow
   ID for the new flow, and MUST include a Timeout option in the DSR
   Options header, giving the lifetime after which state information
   about this flow is to expire.  This packet will generally be a normal
   data packet being sent from this sender to the destination (for
   example, the first packet sent after discovering the new route) but
   is also treated as a "flow establishment" packet.





Johnson, et al.               Experimental                     [Page 21]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   The source node records this flow in its Flow Table for future use,
   setting the TTL in this Flow Table entry to the value used in the TTL
   field in the packet's IP header and setting the Lifetime in this
   entry to the lifetime specified in the Timeout option in the DSR
   Options header.  The TTL field is used for Default Flow Forwarding,
   as described in Sections 3.5.3 and 3.5.4.

   Any further packets sent with this flow ID before the timeout that
   also contain a DSR Options header with a Source Route option MUST use
   this same source route in the Source Route option.

3.5.2.  Receiving and Forwarding Establishment Packets



   Packets intended to establish a flow, as described in Section 3.5.1,
   contain a DSR Options header with a Source Route option and are
   forwarded along the indicated route.  A node implementing the DSR
   flow state extension, when receiving and forwarding such a DSR
   packet, also keeps some state in its own Flow Table to enable it to
   forward future packets that are sent along this flow with only the
   flow ID specified.  Specifically, if the packet also contains a DSR
   Flow State header, this packet SHOULD cause an entry to be
   established for this flow in the Flow Table of each node along the
   packet's route.

   The Hop Count field of the DSR Flow State header is also stored in
   the Flow Table, as is the lifetime specified in the Timeout option
   specified in the DSR Options header.

   If the Flow ID is odd and there is no flow in the Flow Table with
   Flow ID greater than the received Flow ID, set the default Flow ID
   for this (IP Source Address, IP Destination Address) pair to the
   received Flow ID, and the TTL of the packet is recorded.

   The Flow ID option is removed before final delivery of the packet.

3.5.3.  Sending Packets along Established Flows



   When a flow is established as described in Section 3.5.1, a packet is
   sent that establishes state in each node along the route.  This state
   is soft; that is, the protocol contains mechanisms for recovering
   from the loss of this state.  However, the use of these mechanisms
   may result in reduced performance for packets sent along flows with
   forgotten state.  As a result, it is desirable to differentiate
   behavior based on whether or not the sender is reasonably certain
   that the flow state exists on each node along the route.  We define a
   flow's state to be "established end-to-end" if the Flow Tables of all
   nodes on the route contains forwarding information for that flow.
   While it is impossible to detect whether or not a flow's state has



Johnson, et al.               Experimental                     [Page 22]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   been established end-to-end without sending packets, implementations
   may make reasonable assumptions about the retention of flow state and
   the probability that an establishment packet has been seen by all
   nodes on the route.

   A source wishing to send a packet along an established flow
   determines if the flow state has been established end-to-end.  If it
   has not, a DSR Options header with Source Route option with this
   flow's route is added to the packet.  The source SHOULD set the Flow
   ID field of the DSR Flow State header either to the flow ID
   previously associated with this flow's route or to zero.  If it sets
   the Flow ID field to any other value, it MUST follow the processing
   steps in Section 3.5.1 for establishing a new flow ID.  If it sets
   the Flow ID field to a nonzero value, it MUST include a Timeout
   option with a value not greater than the timeout remaining in the
   node's Flow Table, and if its TTL is not equal to that specified in
   the Flow Table, the flow MUST NOT be used as a default flow in the
   future.

   Once flow state has been established end-to-end for non-default
   flows, a source adds a DSR Flow State header to each packet it wishes
   to send along that flow, setting the Flow ID field to the flow ID of
   that flow.  A Source Route option SHOULD NOT be added to the packet,
   though if one is, then the steps for processing flows that have not
   been established end-to-end MUST be followed.

   Once flow state has been established end-to-end for default flows,
   sources sending packets with IP TTL equal to the TTL value in the
   local Flow Table entry for this flow then transmit the packet to the
   next hop.  In this case, a DSR Flow State header SHOULD NOT be added
   to the packet and a DSR Options header likewise SHOULD NOT be added
   to the packet; though if one is, the steps for sending packets along
   non-default flows MUST be followed.  If the IP TTL is not equal to
   the TTL value in the local Flow Table, then the steps for processing
   a non-default flow MUST be followed.

3.5.4.  Receiving and Forwarding Packets Sent along Established Flows



   The handling of packets containing a DSR Options header with both a
   nonzero Flow ID and a Source Route option is described in Section
   3.5.2.  The Flow ID is ignored when it is equal to zero.  This
   section only describes handling of packets without a Source Route
   option.

   If a node receives a packet with a Flow ID in the DSR Options header
   that indicates an unexpired flow in the node's Flow Table, it
   increments the Hop Count in the DSR Options header and forwards the
   packet to the next hop indicated in the Flow Table.



Johnson, et al.               Experimental                     [Page 23]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   If a node receives a packet with a Flow ID that indicates a flow not
   currently in the node's Flow Table, it returns a Route Error of type
   UNKNOWN_FLOW with Error Destination and IP Destination addresses
   copied from the IP Source of the packet triggering the error.  This
   error packet SHOULD be MAC-destined to the node from which the packet
   was received; if it cannot confirm reachability of the previous node
   using Route Maintenance, it MUST send the error as described in
   Section 8.1.1.  The node sending the error SHOULD attempt to salvage
   the packet triggering the Route Error.  If it does salvage the
   packet, it MUST zero the Flow ID in the packet.

   If a node receives a packet with no DSR Options header and no DSR
   Flow State header, it checks the Default Flow Table.  If there is a
   matching entry, it forwards to the next hop indicated in the Flow
   Table for the default flow.  Otherwise, it returns a Route Error of
   type DEFAULT_FLOW_UNKNOWN with Error Destination and IP Destination
   addresses copied from the IP Source Address of the packet triggering
   the error.  This error packet SHOULD be MAC-destined to the node from
   which it was received; if this node cannot confirm reachability of
   the previous node using Route Maintenance, it MUST send the error as
   described in Section 8.1.1.  The node sending the error SHOULD
   attempt to salvage the packet triggering the Route Error.  If it does
   salvage the packet, it MUST zero the Flow ID in the packet.

3.5.5.  Processing Route Errors



   When a node receives a Route Error of type UNKNOWN_FLOW, it marks the
   flow to indicate that it has not been established end-to-end.  When a
   node receives a Route Error of type DEFAULT_FLOW_UNKNOWN, it marks
   the default flow to indicate that it has not been established end-
   to-end.

3.5.6.  Interaction with Automatic Route Shortening



   Because a full source route is not carried in every packet, an
   alternative method for performing automatic route shortening is
   necessary for packets using the flow state extension.  Instead, nodes
   promiscuously listen to packets, and if a node receives a packet with
   (IP Source, IP Destination, Flow ID) found in the Flow Table but the
   MAC-layer (next hop) destination address of the packet is not this
   node, the node determines whether the packet was sent by an upstream
   or downstream node by examining the Hop Count field in the DSR Flow
   State header.  If the Hop Count field is less than the expected Hop
   Count at this node (that is, the expected Hop Count field in the Flow
   Table described in Section 5.1), the node assumes that the packet was
   sent by an upstream node and adds an entry for the packet to its
   Automatic Route Shortening Table, possibly evicting an earlier entry
   added to this table.  When the packet is then sent to that node for



Johnson, et al.               Experimental                     [Page 24]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   forwarding, the node finds that it has previously received the packet
   by checking its Automatic Route Shortening Table and returns a
   gratuitous Route Reply to the source of the packet.

3.5.7.  Loop Detection



   If a node receives a packet for forwarding with TTL lower than
   expected and default flow forwarding is being used, it sends a Route
   Error of type DEFAULT_FLOW_UNKNOWN back to the IP source.  It can
   attempt delivery of the packet by normal salvaging (subject to
   constraints described in Section 8.6.7).

3.5.8.  Acknowledgement Destination



   In packets sent using Flow State, the previous hop is not necessarily
   known.  In order to allow nodes that have lost flow state to
   determine the previous hop, the address of the previous hop can
   optionally be stored in the Acknowledgement Request.  This extension
   SHOULD NOT be used when a Source Route option is present, MAY be used
   when flow state routing is used without a Source Route option, and
   SHOULD be used before Route Maintenance determines that the next-hop
   destination is unreachable.

3.5.9.  Crash Recovery



   Each node has a maximum Timeout value that it can possibly generate.
   This can be based on the largest number that can be set in a timeout
   option (2**16 - 1 seconds) or may be less than this, set in system
   software.  When a node crashes, it does not establish new flows for a
   period equal to this maximum Timeout value, in order to avoid
   colliding with its old Flow IDs.

3.5.10.  Rate Limiting



   Flow IDs can be assigned with a counter.  More specifically, the
   "Current Flow ID" is kept.  When a new default Flow ID needs to be
   assigned, if the Current Flow ID is odd, the Current Flow ID is
   assigned as the Flow ID and the Current Flow ID is incremented by
   one; if the Current Flow ID is even, one plus the Current Flow ID is
   assigned as the Flow ID and the Current Flow ID is incremented by
   two.

   If Flow IDs are assigned in this way, one algorithm for avoiding
   duplicate, unexpired Flow IDs is to rate limit new Flow IDs to an
   average rate of n assignments per second, where n is 2**15 divided by
   the maximum Timeout value.  This can be averaged over any period not
   exceeding the maximum Timeout value.




Johnson, et al.               Experimental                     [Page 25]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


3.5.11.  Interaction with Packet Salvaging



   Salvaging is modified to zero the Flow ID field in the packet.  Also,
   anytime this document refers to the Salvage field in the Source Route
   option in a DSR Options header, packets without a Source Route option
   are considered to have the value zero in the Salvage field.

4.  Conceptual Data Structures



   This document describes the operation of the DSR protocol in terms of
   a number of conceptual data structures.  This section describes each
   of these data structures and provides an overview of its use in the
   protocol.  In an implementation of the protocol, these data
   structures MUST be implemented in a manner consistent with the
   external behavior described in this document, but the choice of
   implementation used is otherwise unconstrained.  Additional
   conceptual data structures are required for the optional flow state
   extensions to DSR; these data structures are described in Section 5.

4.1.  Route Cache



   Each node implementing DSR MUST maintain a Route Cache, containing
   routing information needed by the node.  A node adds information to
   its Route Cache as it learns of new links between nodes in the ad hoc
   network; for example, a node may learn of new links when it receives
   a packet carrying a Route Request, Route Reply, or DSR source route.
   Likewise, a node removes information from its Route Cache as it
   learns that existing links in the ad hoc network have broken.  For
   example, a node may learn of a broken link when it receives a packet
   carrying a Route Error or through the link-layer retransmission
   mechanism reporting a failure in forwarding a packet to its next-hop
   destination.

   Anytime a node adds new information to its Route Cache, the node
   SHOULD check each packet in its own Send Buffer (Section 4.2) to
   determine whether a route to that packet's IP Destination Address now
   exists in the node's Route Cache (including the information just
   added to the Cache).  If so, the packet SHOULD then be sent using
   that route and removed from the Send Buffer.

   It is possible to interface a DSR network with other networks,
   external to this DSR network.  Such external networks may, for
   example, be the Internet or may be other ad hoc networks routed with
   a routing protocol other than DSR.  Such external networks may also
   be other DSR networks that are treated as external networks in order
   to improve scalability.  The complete handling of such external
   networks is beyond the scope of this document.  However, this
   document specifies a minimal set of requirements and features



Johnson, et al.               Experimental                     [Page 26]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   necessary to allow nodes only implementing this specification to
   interoperate correctly with nodes implementing interfaces to such
   external networks.  This minimal set of requirements and features
   involve the First Hop External (F) and Last Hop External (L) bits in
   a DSR Source Route option (Section 6.7) and a Route Reply option
   (Section 6.3) in a packet's DSR Options header (Section 6).  These
   requirements also include the addition of an External flag bit
   tagging each link in the Route Cache, copied from the First Hop
   External (F) and Last Hop External (L) bits in the DSR Source Route
   option or Route Reply option from which this link was learned.

   The Route Cache SHOULD support storing more than one route to each
   destination.  In searching the Route Cache for a route to some
   destination node, the Route Cache is searched by destination node
   address.  The following properties describe this searching function
   on a Route Cache:

   -  Each implementation of DSR at any node MAY choose any appropriate
      strategy and algorithm for searching its Route Cache and selecting
      a "best" route to the destination from among those found.  For
      example, a node MAY choose to select the shortest route to the
      destination (the shortest sequence of hops), or it MAY use an
      alternate metric to select the route from the Cache.

   -  However, if there are multiple cached routes to a destination, the
      selection of routes when searching the Route Cache SHOULD prefer
      routes that do not have the External flag set on any link.  This
      preference will select routes that lead directly to the target
      node over routes that attempt to reach the target via any external
      networks connected to the DSR ad hoc network.

   -  In addition, any route selected when searching the Route Cache
      MUST NOT have the External bit set for any links other than
      possibly the first link, the last link, or both; the External bit
      MUST NOT be set for any intermediate hops in the route selected.

   An implementation of a Route Cache MAY provide a fixed capacity for
   the cache, or the cache size MAY be variable.  The following
   properties describe the management of available space within a node's
   Route Cache:

   -  Each implementation of DSR at each node MAY choose any appropriate
      policy for managing the entries in its Route Cache, such as when
      limited cache capacity requires a choice of which entries to
      retain in the Cache.  For example, a node MAY chose a "least
      recently used" (LRU) cache replacement policy, in which the entry





Johnson, et al.               Experimental                     [Page 27]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      last used longest ago is discarded from the cache if a decision
      needs to be made to allow space in the cache for some new entry
      being added.

   -  However, the Route Cache replacement policy SHOULD allow routes to
      be categorized based upon "preference", where routes with a higher
      preferences are less likely to be removed from the cache.  For
      example, a node could prefer routes for which it initiated a Route
      Discovery over routes that it learned as the result of promiscuous
      snooping on other packets.  In particular, a node SHOULD prefer
      routes that it is presently using over those that it is not.

   Any suitable data structure organization, consistent with this
   specification, MAY be used to implement the Route Cache in any node.
   For example, the following two types of organization are possible:

   -  In DSR, the route returned in each Route Reply that is received by
      the initiator of a Route Discovery (or that is learned from the
      header of overhead packets, as described in Section 8.1.4)
      represents a complete path (a sequence of links) leading to the
      destination node.  By caching each of these paths separately, a
      "path cache" organization for the Route Cache can be formed.  A
      path cache is very simple to implement and easily guarantees that
      all routes are loop-free, since each individual route from a Route
      Reply or Route Request or used in a packet is loop-free.  To
      search for a route in a path cache data structure, the sending
      node can simply search its Route Cache for any path (or prefix of
      a path) that leads to the intended destination node.

      This type of organization for the Route Cache in DSR has been
      extensively studied through simulation [BROCH98, HU00,
      JOHANSSON99, MALTZ99a] and through implementation of DSR in a
      mobile outdoor testbed under significant workload [MALTZ99b,
      MALTZ00, MALTZ01].

   -  Alternatively, a "link cache" organization could be used for the
      Route Cache, in which each individual link (hop) in the routes
      returned in Route Reply packets (or otherwise learned from the
      header of overhead packets) is added to a unified graph data
      structure of this node's current view of the network topology.  To
      search for a route in link cache, the sending node must use a more
      complex graph search algorithm, such as the well-known Dijkstra's
      shortest-path algorithm, to find the current best path through the
      graph to the destination node.  Such an algorithm is more
      difficult to implement and may require significantly more CPU time
      to execute.





Johnson, et al.               Experimental                     [Page 28]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      However, a link cache organization is more powerful than a path
      cache organization, in its ability to effectively utilize all of
      the potential information that a node might learn about the state
      of the network.  In particular, links learned from different Route
      Discoveries or from the header of any overheard packets can be
      merged together to form new routes in the network, but this is not
      possible in a path cache due to the separation of each individual
      path in the cache.

      This type of organization for the Route Cache in DSR, including
      the effect of a range of implementation choices, has been studied
      through detailed simulation [HU00].

   The choice of data structure organization to use for the Route Cache
   in any DSR implementation is a local matter for each node and affects
   only performance; any reasonable choice of organization for the Route
   Cache does not affect either correctness or interoperability.

   Each entry in the Route Cache SHOULD have a timeout associated with
   it, to allow that entry to be deleted if not used within some time.
   The particular choice of algorithm and data structure used to
   implement the Route Cache SHOULD be considered in choosing the
   timeout for entries in the Route Cache.  The configuration variable
   RouteCacheTimeout defined in Section 9 specifies the timeout to be
   applied to entries in the Route Cache, although it is also possible
   to instead use an adaptive policy in choosing timeout values rather
   than using a single timeout setting for all entries.  For example,
   the Link-MaxLife cache design (below) uses an adaptive timeout
   algorithm and does not use the RouteCacheTimeout configuration
   variable.

   As guidance to implementers, Appendix A describes a type of link
   cache known as "Link-MaxLife" that has been shown to outperform other
   types of link caches and path caches studied in detailed simulation
   [HU00].  Link-MaxLife is an adaptive link cache in which each link in
   the cache has a timeout that is determined dynamically by the caching
   node according to its observed past behavior of the two nodes at the
   ends of the link.  In addition, when selecting a route for a packet
   being sent to some destination, among cached routes of equal length
   (number of hops) to that destination, Link-MaxLife selects the route
   with the longest expected lifetime (highest minimum timeout of any
   link in the route).  Use of the Link-MaxLife design for the Route
   Cache is recommended in implementations of DSR.








Johnson, et al.               Experimental                     [Page 29]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


4.2.  Send Buffer



   The Send Buffer of a node implementing DSR is a queue of packets that
   cannot be sent by that node because it does not yet have a source
   route to each such packet's destination.  Each packet in the Send
   Buffer is logically associated with the time that it was placed into
   the buffer and SHOULD be removed from the Send Buffer and silently
   discarded after a period of SendBufferTimeout after initially being
   placed in the buffer.  If necessary, a FIFO strategy SHOULD be used
   to evict packets before they time out to prevent the buffer from
   overflowing.

   Subject to the rate limiting defined in Section 4.3, a Route
   Discovery SHOULD be initiated as often as allowed for the destination
   address of any packets residing in the Send Buffer.

4.3.  Route Request Table



   The Route Request Table of a node implementing DSR records
   information about Route Requests that have been recently originated
   or forwarded by this node.  The table is indexed by IP address.

   The Route Request Table on a node records the following information
   about nodes to which this node has initiated a Route Request:

   -  The Time-to-Live (TTL) field used in the IP header of the Route
      Request for the last Route Discovery initiated by this node for
      that target node.  This value allows the node to implement a
      variety of algorithms for controlling the spread of its Route
      Request on each Route Discovery initiated for a target.  As
      examples, two possible algorithms for this use of the TTL field
      are described in Section 3.3.3.

   -  The time that this node last originated a Route Request for that
      target node.

   -  The number of consecutive Route Discoveries initiated for this
      target since receiving a valid Route Reply giving a route to that
      target node.

   -  The remaining amount of time before which this node MAY next
      attempt at a Route Discovery for that target node.  When the node
      initiates a new Route Discovery for this target node, this field
      in the Route Request Table entry for that target node is
      initialized to the timeout for that Route Discovery, after which
      the node MAY initiate a new Discovery for that target.  Until a
      valid Route Reply is received for this target node address, a node
      MUST implement a back-off algorithm in determining this timeout



Johnson, et al.               Experimental                     [Page 30]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      value for each successive Route Discovery initiated for this
      target using the same Time-to-Live (TTL) value in the IP header of
      the Route Request packet.  The timeout between such consecutive
      Route Discovery initiations SHOULD increase by doubling the
      timeout value on each new initiation.

   In addition, the Route Request Table on a node also records the
   following information about initiator nodes from which this node has
   received a Route Request:

   -  A FIFO cache of size RequestTableIds entries containing the
      Identification value and target address from the most recent Route
      Requests received by this node from that initiator node.

   Nodes SHOULD use an LRU policy to manage the entries in their Route
   Request Table.

   The number of Identification values to retain in each Route Request
   Table entry, RequestTableIds, MUST NOT be unlimited, since, in the
   worst case, when a node crashes and reboots, the first
   RequestTableIds Route Discoveries it initiates after rebooting could
   appear to be duplicates to the other nodes in the network.  In
   addition, a node SHOULD base its initial Identification value, used
   for Route Discoveries after rebooting, on a battery backed-up clock
   or other persistent memory device, if available, in order to help
   avoid any possible such delay in successfully discovering new routes
   after rebooting; if no such source of initial Identification value is
   available, a node after rebooting SHOULD base its initial
   Identification value on a random number.

4.4.  Gratuitous Route Reply Table



   The Gratuitous Route Reply Table of a node implementing DSR records
   information about "gratuitous" Route Replies sent by this node as
   part of automatic route shortening.  As described in Section 3.4.3, a
   node returns a gratuitous Route Reply when it overhears a packet
   transmitted by some node, for which the node overhearing the packet
   was not the intended next-hop node but was named later in the
   unexpended hops of the source route in that packet; the node
   overhearing the packet returns a gratuitous Route Reply to the
   original sender of the packet, listing the shorter route (not
   including the hops of the source route "skipped over" by this
   packet).  A node uses its Gratuitous Route Reply Table to limit the
   rate at which it originates gratuitous Route Replies to the same
   original sender for the same node from which it overheard a packet to
   trigger the gratuitous Route Reply.





Johnson, et al.               Experimental                     [Page 31]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   Each entry in the Gratuitous Route Reply Table of a node contains the
   following fields:

   -  The address of the node to which this node originated a gratuitous
      Route Reply.

   -  The address of the node from which this node overheard the packet
      triggering that gratuitous Route Reply.

   -  The remaining time before which this entry in the Gratuitous Route
      Reply Table expires and SHOULD be deleted by the node.  When a
      node creates a new entry in its Gratuitous Route Reply Table, the
      timeout value for that entry SHOULD be initialized to the value
      GratReplyHoldoff.

   When a node overhears a packet that would trigger a gratuitous Route
   Reply, if a corresponding entry already exists in the node's
   Gratuitous Route Reply Table, then the node SHOULD NOT send a
   gratuitous Route Reply for that packet.  Otherwise (i.e., if no
   corresponding entry already exists), the node SHOULD create a new
   entry in its Gratuitous Route Reply Table to record that gratuitous
   Route Reply, with a timeout value of GratReplyHoldoff.

4.5.  Network Interface Queue and Maintenance Buffer



   Depending on factors such as the structure and organization of the
   operating system, protocol stack implementation, network interface
   device driver, and network interface hardware, a packet being
   transmitted could be queued in a variety of ways.  For example,
   outgoing packets from the network protocol stack might be queued at
   the operating system or link layer, before transmission by the
   network interface.  The network interface might also provide a
   retransmission mechanism for packets, such as occurs in IEEE 802.11
   [IEEE80211]; the DSR protocol, as part of Route Maintenance, requires
   limited buffering of packets already transmitted for which the
   reachability of the next-hop destination has not yet been determined.
   The operation of DSR is defined here in terms of two conceptual data
   structures that, together, incorporate this queuing behavior.

   The Network Interface Queue of a node implementing DSR is an output
   queue of packets from the network protocol stack waiting to be
   transmitted by the network interface; for example, in the 4.4BSD Unix
   network protocol stack implementation, this queue for a network
   interface is represented as a "struct ifqueue" [WRIGHT95].  This
   queue is used to hold packets while the network interface is in the
   process of transmitting another packet.





Johnson, et al.               Experimental                     [Page 32]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   The Maintenance Buffer of a node implementing DSR is a queue of
   packets sent by this node that are awaiting next-hop reachability
   confirmation as part of Route Maintenance.  For each packet in the
   Maintenance Buffer, a node maintains a count of the number of
   retransmissions and the time of the last retransmission.  Packets are
   added to the Maintenance buffer after the first transmission attempt
   is made.  The Maintenance Buffer MAY be of limited size; when adding
   a new packet to the Maintenance Buffer, if the buffer size is
   insufficient to hold the new packet, the new packet SHOULD be
   silently discarded.  If, after MaxMaintRexmt attempts to confirm
   next-hop reachability of some node, no confirmation is received, all
   packets in this node's Maintenance Buffer with this next-hop
   destination SHOULD be removed from the Maintenance Buffer.  In this
   case, the node also SHOULD originate a Route Error for this packet to
   each original source of a packet removed in this way (Section 8.3)
   and SHOULD salvage each packet removed in this way (Section 8.3.6) if
   it has another route to that packet's IP Destination Address in its
   Route Cache.  The definition of MaxMaintRexmt conceptually includes
   any retransmissions that might be attempted for a packet at the link
   layer or within the network interface hardware.  The timeout value to
   use for each transmission attempt for an acknowledgement request
   depends on the type of acknowledgement mechanism used by Route
   Maintenance for that attempt, as described in Section 8.3.

4.6.  Blacklist



   When a node using the DSR protocol is connected through a network
   interface that requires physically bidirectional links for unicast
   transmission, the node MUST maintain a blacklist.  The blacklist is a
   table, indexed by neighbor node address, that indicates that the link
   between this node and the specified neighbor node may not be
   bidirectional.  A node places another node's address in this list
   when it believes that broadcast packets from that other node reach
   this node, but that unicast transmission between the two nodes is not
   possible.  For example, if a node forwarding a Route Reply discovers
   that the next hop is unreachable, it places that next hop in the
   node's blacklist.

   Once a node discovers that it can communicate bidirectionally with
   one of the nodes listed in the blacklist, it SHOULD remove that node
   from the blacklist.  For example, if node A has node B listed in its
   blacklist, but after transmitting a Route Request, node A hears B
   forward the Route Request with a route record indicating that the
   broadcast from A to B was successful, then A SHOULD remove the entry
   for node B from its blacklist.






Johnson, et al.               Experimental                     [Page 33]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   A node MUST associate a state with each node listed in its blacklist,
   specifying whether the unidirectionality of the link to that node is
   "questionable" or "probable".  Each time the unreachability is
   positively determined, the node SHOULD set the state to "probable".
   After the unreachability has not been positively determined for some
   amount of time, the state SHOULD revert to "questionable".  A node
   MAY expire entries for nodes from its blacklist after a reasonable
   amount of time.

5.  Additional Conceptual Data Structures for Flow State Extension



   This section defines additional conceptual data structures used by
   the optional "flow state" extension to DSR.  In an implementation of
   the protocol, these data structures MUST be implemented in a manner
   consistent with the external behavior described in this document, but
   the choice of implementation used is otherwise unconstrained.

5.1.  Flow Table



   A node implementing the flow state extension MUST implement a Flow
   Table or other data structure consistent with the external behavior
   described in this section.  A node not implementing the flow state
   extension SHOULD NOT implement a Flow Table.

   The Flow Table records information about flows from which packets
   recently have been sent or forwarded by this node.  The table is
   indexed by a triple (IP Source Address, IP Destination Address, Flow
   ID), where Flow ID is a 16-bit number assigned by the source as
   described in Section 3.5.1.  Each entry in the Flow Table contains
   the following fields:

   -  The MAC address of the next-hop node along this flow.

   -  An indication of the outgoing network interface on this node to be
      used in transmitting packets along this flow.

   -  The MAC address of the previous-hop node along this flow.

   -  An indication of the network interface on this node from which
      packets from that previous-hop node are received.

   -  A timeout after which this entry in the Flow Table MUST be
      deleted.

   -  The expected value of the Hop Count field in the DSR Flow State
      header for packets received for forwarding along this field (for
      use with packets containing a DSR Flow State header).




Johnson, et al.               Experimental                     [Page 34]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  An indication of whether or not this flow can be used as a default
      flow for packets originated by this node (the Flow ID of a default
      flow MUST be odd).

   -  The entry SHOULD record the complete source route for the flow.
      (Nodes not recording the complete source route cannot participate
      in Automatic Route Shortening.)

   -  The entry MAY contain a field recording the time this entry was
      last used.

   The entry MUST be deleted when its timeout expires.

5.2.  Automatic Route Shortening Table



   A node implementing the flow state extension SHOULD implement an
   Automatic Route Shortening Table or other data structure consistent
   with the external behavior described in this section.  A node not
   implementing the flow state extension SHOULD NOT implement an
   Automatic Route Shortening Table.

   The Automatic Route Shortening Table records information about
   received packets for which Automatic Route Shortening may be
   possible.  The table is indexed by a triple (IP Source Address, IP
   Destination Address, Flow ID).  Each entry in the Automatic Route
   Shortening Table contains a list of (packet identifier, Hop Count)
   pairs for that flow.  The packet identifier in the list may be any
   unique identifier for the received packet; for example, for IPv4
   packets, the combination of the following fields from the packet's IP
   header MAY be used as a unique identifier for the packet:  Source
   Address, Destination Address, Identification, Protocol, Fragment
   Offset, and Total Length.  The Hop Count in the list in the entry is
   copied from the Hop Count field in the DSR Flow State header of the
   received packet for which this table entry was created.  Any packet
   identifier SHOULD appear at most once in an entry's list, and this
   list item SHOULD record the minimum Hop Count value received for that
   packet (if the wireless signal strength or signal-to-noise ratio at
   which a packet is received is available to the DSR implementation in
   a node, the node MAY, for example, remember instead in this list the
   minimum Hop Count value for which the received packet's signal
   strength or signal-to-noise ratio exceeded some threshold).

   Space in the Automatic Route Shortening Table of a node MAY be
   dynamically managed by any local algorithm at the node.  For example,
   in order to limit the amount of memory used to store the table, any
   existing entry MAY be deleted at any time, and the number of packets
   listed in each entry MAY be limited.  However, when reclaiming space
   in the table, nodes SHOULD favor retaining information about more



Johnson, et al.               Experimental                     [Page 35]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   flows in the table rather than about more packets listed in each
   entry in the table, as long as at least the listing of some small
   number of packets (e.g., 3) can be retained in each entry.

5.3.  Default Flow ID Table



   A node implementing the flow state extension MUST implement a Default
   Flow Table or other data structure consistent with the external
   behavior described in this section.  A node not implementing the flow
   state extension SHOULD NOT implement a Default Flow Table.

   For each (IP Source Address, IP Destination Address) pair for which a
   node forwards packets, the node MUST record:

   -  The largest odd Flow ID value seen.

   -  The time at which all the corresponding flows that are forwarded
      by this node expire.

   -  The current default Flow ID.

   -  A flag indicating whether or not the current default Flow ID is
      valid.

   If a node deletes this record for an (IP Source Address, IP
   Destination Address) pair, it MUST also delete all Flow Table entries
   for that pair.  Nodes MUST delete table entries if all of this (IP
   Source Address, IP Destination Address) pair's flows that are
   forwarded by this node expire.

6.  DSR Options Header Format



   The Dynamic Source Routing protocol makes use of a special header
   carrying control information that can be included in any existing IP
   packet.  This DSR Options header in a packet contains a small fixed-
   sized, 4-octet portion, followed by a sequence of zero or more DSR
   options carrying optional information.  The end of the sequence of
   DSR options in the DSR Options header is implied by the total length
   of the DSR Options header.

   For IPv4, the DSR Options header MUST immediately follow the IP
   header in the packet.  (If a Hop-by-Hop Options extension header, as
   defined in IPv6 [RFC2460], becomes defined for IPv4, the DSR Options
   header MUST immediately follow the Hop-by-Hop Options extension
   header, if one is present in the packet, and MUST otherwise
   immediately follow the IP header.)





Johnson, et al.               Experimental                     [Page 36]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   To add a DSR Options header to a packet, the DSR Options header is
   inserted following the packet's IP header, before any following
   header such as a traditional (e.g., TCP or UDP) transport layer
   header.  Specifically, the Protocol field in the IP header is used to
   indicate that a DSR Options header follows the IP header, and the
   Next Header field in the DSR Options header is used to indicate the
   type of protocol header (such as a transport layer header) following
   the DSR Options header.

   If any headers follow the DSR Options header in a packet, the total
   length of the DSR Options header (and thus the total, combined length
   of all DSR options present) MUST be a multiple of 4 octets.  This
   requirement preserves the alignment of these following headers in the
   packet.

6.1.  Fixed Portion of DSR Options Header



   The fixed portion of the DSR Options header is used to carry
   information that must be present in any DSR Options header.  This
   fixed portion of the DSR Options header has the following format:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Next Header  |F|   Reserved  |        Payload Length         |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   .                                                               .
   .                            Options                            .
   .                                                               .
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Next Header

         8-bit selector.  Identifies the type of header immediately
         following the DSR Options header.  Uses the same values as the
         IPv4 Protocol field [RFC1700].  If no header follows, then Next
         Header MUST have the value 59, "No Next Header" [RFC2460].

      Flow State Header (F)

         Flag bit.  MUST be set to 0.  This bit is set in a DSR Flow
         State header (Section 7.1) and clear in a DSR Options header.

      Reserved

         MUST be sent as 0 and ignored on reception.





Johnson, et al.               Experimental                     [Page 37]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Payload Length

         The length of the DSR Options header, excluding the 4-octet
         fixed portion.  The value of the Payload Length field defines
         the total length of all options carried in the DSR Options
         header.

      Options

         Variable-length field; the length of the Options field is
         specified by the Payload Length field in this DSR Options
         header.  Contains one or more pieces of optional information
         (DSR options), encoded in type-length-value (TLV) format (with
         the exception of the Pad1 option described in Section 6.8).

   The placement of DSR options following the fixed portion of the DSR
   Options header MAY be padded for alignment.  However, due to the
   typically limited available wireless bandwidth in ad hoc networks,
   this padding is not required, and receiving nodes MUST NOT expect
   options within a DSR Options header to be aligned.

   Each DSR option is assigned a unique Option Type code.  The most
   significant 3 bits (that is, Option Type & 0xE0) allow a node not
   implementing processing for this Option Type value to behave in the
   manner closest to correct for that type:

   -  The most significant bit in the Option Type value (that is, Option
      Type & 0x80) represents whether or not a node receiving this
      Option Type (when the node does not implement processing for this
      Option Type) SHOULD respond to such a DSR option with a Route
      Error of type OPTION_NOT_SUPPORTED, except that such a Route Error
      SHOULD never be sent in response to a packet containing a Route
      Request option.

   -  The two following bits in the Option Type value (that is, Option
      Type & 0x60) are a two-bit field indicating how such a node that
      does not support this Option Type MUST process the packet:

         00 = Ignore Option
         01 = Remove Option
         10 = Mark Option
         11 = Drop Packet



      When these 2 bits are 00 (that is, Option Type & 0x60 == 0), a
      node not implementing processing for that Option Type MUST use the
      Opt Data Len field to skip over the option and continue
      processing.  When these 2 bits are 01 (that is, Option Type & 0x60
      == 0x20), a node not implementing processing for that Option Type



Johnson, et al.               Experimental                     [Page 38]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      MUST use the Opt Data Len field to remove the option from the
      packet and continue processing as if the option had not been
      included in the received packet.  When these 2 bits are 10 (that
      is, Option Type & 0x60 == 0x40), a node not implementing
      processing for that Option Type MUST set the most significant bit
      following the Opt Data Len field, MUST ignore the contents of the
      option using the Opt Data Len field, and MUST continue processing
      the packet.  Finally, when these 2 bits are 11 (that is, Option
      Type & 0x60 == 0x60), a node not implementing processing for that
      Option Type MUST drop the packet.

   The following types of DSR options are defined in this document for
   use within a DSR Options header:

   -  Route Request option (Section 6.2)

   -  Route Reply option (Section 6.3)

   -  Route Error option (Section 6.4)

   -  Acknowledgement Request option (Section 6.5)

   -  Acknowledgement option (Section 6.6)

   -  DSR Source Route option (Section 6.7)

   -  Pad1 option (Section 6.8)

   -  PadN option (Section 6.9)

   In addition, Section 7 specifies further DSR options for use with the
   optional DSR flow state extension.



















Johnson, et al.               Experimental                     [Page 39]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


6.2.  Route Request Option



   The Route Request option in a DSR Options header is encoded as
   follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  |  Opt Data Len |         Identification        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                         Target Address                        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[1]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[2]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                              ...                              |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[n]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   IP fields:

      Source Address

         MUST be set to the address of the node originating this packet.
         Intermediate nodes that retransmit the packet to propagate the
         Route Request MUST NOT change this field.

      Destination Address

         MUST be set to the IP limited broadcast address
         (255.255.255.255).

      Hop Limit (TTL)

         MAY be varied from 1 to 255, for example, to implement non-
         propagating Route Requests and Route Request expanding-ring
         searches (Section 3.3.3).

   Route Request fields:

      Option Type

         1.  Nodes not understanding this option will ignore this
         option.





Johnson, et al.               Experimental                     [Page 40]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.  MUST be set
         equal to (4 * n) + 6, where n is the number of addresses in the
         Route Request Option.

      Identification

         A unique value generated by the initiator (original sender) of
         the Route Request.  Nodes initiating a Route Request generate a
         new Identification value for each Route Request, for example
         based on a sequence number counter of all Route Requests
         initiated by the node.

         This value allows a receiving node to determine whether it has
         recently seen a copy of this Route Request.  If this
         Identification value (for this IP Source address and Target
         Address) is found by this receiving node in its Route Request
         Table (in the cache of Identification values in the entry there
         for this initiating node), this receiving node MUST discard the
         Route Request.  When a Route Request is propagated, this field
         MUST be copied from the received copy of the Route Request
         being propagated.

      Target Address

         The address of the node that is the target of the Route
         Request.

      Address[1..n]

         Address[i] is the IPv4 address of the i-th node recorded in the
         Route Request option.  The address given in the Source Address
         field in the IP header is the address of the initiator of the
         Route Discovery and MUST NOT be listed in the Address[i]
         fields; the address given in Address[1] is thus the IPv4
         address of the first node on the path after the initiator.  The
         number of addresses present in this field is indicated by the
         Opt Data Len field in the option (n = (Opt Data Len - 6) / 4).
         Each node propagating the Route Request adds its own address to
         this list, increasing the Opt Data Len value by 4 octets.

   The Route Request option MUST NOT appear more than once within a DSR
   Options header.






Johnson, et al.               Experimental                     [Page 41]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


6.3.  Route Reply Option



   The Route Reply option in a DSR Options header is encoded as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
                   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
                   |  Option Type  |  Opt Data Len |L|   Reserved  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[1]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[2]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                              ...                              |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[n]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   IP fields:

      Source Address

         Set to the address of the node sending the Route Reply.  In the
         case of a node sending a reply from its Route Cache (Section
         3.3.2) or sending a gratuitous Route Reply (Section 3.4.3),
         this address can differ from the address that was the target of
         the Route Discovery.

      Destination Address

         MUST be set to the address of the source node of the route
         being returned.  Copied from the Source Address field of the
         Route Request generating the Route Reply or, in the case of a
         gratuitous Route Reply, copied from the Source Address field of
         the data packet triggering the gratuitous Reply.

   Route Reply fields:

      Option Type

         2.  Nodes not understanding this option will ignore this
         option.









Johnson, et al.               Experimental                     [Page 42]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.  MUST be set
         equal to (4 * n) + 1, where n is the number of addresses in the
         Route Reply Option.

      Last Hop External (L)

         Set to indicate that the last hop given by the Route Reply (the
         link from Address[n-1] to Address[n]) is actually an arbitrary
         path in a network external to the DSR network; the exact route
         outside the DSR network is not represented in the Route Reply.
         Nodes caching this hop in their Route Cache MUST flag the
         cached hop with the External flag.  Such hops MUST NOT be
         returned in a cached Route Reply generated from this Route
         Cache entry, and selection of routes from the Route Cache to
         route a packet being sent SHOULD prefer routes that contain no
         hops flagged as External.

      Reserved

         MUST be sent as 0 and ignored on reception.

      Address[1..n]

         The source route being returned by the Route Reply.  The route
         indicates a sequence of hops, originating at the source node
         specified in the Destination Address field of the IP header of
         the packet carrying the Route Reply, through each of the
         Address[i] nodes in the order listed in the Route Reply, ending
         at the node indicated by Address[n].  The number of addresses
         present in the Address[1..n] field is indicated by the Opt Data
         Len field in the option (n = (Opt Data Len - 1) / 4).

   A Route Reply option MAY appear one or more times within a DSR
   Options header.














Johnson, et al.               Experimental                     [Page 43]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


6.4.  Route Error Option



   The Route Error option in a DSR Options header is encoded as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  |  Opt Data Len |   Error Type  |Reservd|Salvage|
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                      Error Source Address                     |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                   Error Destination Address                   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   .                                                               .
   .                   Type-Specific Information                   .
   .                                                               .
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Option Type

         3.  Nodes not understanding this option will ignore this
         option.

      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

         For the current definition of the Route Error option,
         this field MUST be set to 10, plus the size of any
         Type-Specific Information present in the Route Error.  Further
         extensions to the Route Error option format may also be
         included after the Type-Specific Information portion of the
         Route Error option specified above.  The presence of such
         extensions will be indicated by the Opt Data Len field.
         When the Opt Data Len is greater than that required for
         the fixed portion of the Route Error plus the necessary
         Type-Specific Information as indicated by the Option Type
         value in the option, the remaining octets are interpreted as
         extensions.  Currently, no such further extensions have been
         defined.

      Error Type

         The type of error encountered.  Currently, the following type
         values are defined:





Johnson, et al.               Experimental                     [Page 44]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


            1 = NODE_UNREACHABLE
            2 = FLOW_STATE_NOT_SUPPORTED
            3 = OPTION_NOT_SUPPORTED

         Other values of the Error Type field are reserved for future
         use.

      Reservd

         Reserved.  MUST be sent as 0 and ignored on reception.

      Salvage

         A 4-bit unsigned integer.  Copied from the Salvage field in the
         DSR Source Route option of the packet triggering the Route
         Error.

         The "total salvage count" of the Route Error option is derived
         from the value in the Salvage field of this Route Error option
         and all preceding Route Error options in the packet as follows:
         the total salvage count is the sum of, for each such Route
         Error option, one plus the value in the Salvage field of that
         Route Error option.

      Error Source Address

         The address of the node originating the Route Error (e.g., the
         node that attempted to forward a packet and discovered the link
         failure).

      Error Destination Address

         The address of the node to which the Route Error must be
         delivered.  For example, when the Error Type field is set to
         NODE_UNREACHABLE, this field will be set to the address of the
         node that generated the routing information claiming that the
         hop from the Error Source Address to Unreachable Node Address
         (specified in the Type-Specific Information) was a valid hop.

      Type-Specific Information

         Information specific to the Error Type of this Route Error
         message.

   A Route Error option MAY appear one or more times within a DSR
   Options header.





Johnson, et al.               Experimental                     [Page 45]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


6.4.1.  Node Unreachable Type-Specific Information



   When the Route Error is of type NODE_UNREACHABLE, the Type-Specific
   Information field is defined as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                    Unreachable Node Address                   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Unreachable Node Address

         The IP address of the node that was found to be unreachable
         (the next-hop neighbor to which the node with address
         Error Source Address was attempting to transmit the packet).

6.4.2.  Flow State Not Supported Type-Specific Information



   When the Route Error is of type FLOW_STATE_NOT_SUPPORTED, the
   Type-Specific Information field is empty.

6.4.3.  Option Not Supported Type-Specific Information



   When the Route Error is of type OPTION_NOT_SUPPORTED, the
   Type-Specific Information field is defined as follows:

   0 1 2 3 4 5 6 7
   +-+-+-+-+-+-+-+-+
   |Unsupported Opt|
   +-+-+-+-+-+-+-+-+



      Unsupported Opt

         The Option Type of option triggering the Route Error.

6.5.  Acknowledgement Request Option



   The Acknowledgement Request option in a DSR Options header is encoded
   as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  |  Opt Data Len |         Identification        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+





Johnson, et al.               Experimental                     [Page 46]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Option Type

         160.  Nodes not understanding this option will remove the
         option and return a Route Error.



      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

      Identification

         The Identification field is set to a unique value and is copied
         into the Identification field of the Acknowledgement option
         when returned by the node receiving the packet over this hop.

   An Acknowledgement Request option MUST NOT appear more than once
   within a DSR Options header.

6.6.  Acknowledgement Option



   The Acknowledgement option in a DSR Options header is encoded as
   follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  |  Opt Data Len |         Identification        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                       ACK Source Address                      |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                     ACK Destination Address                   |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Option Type

         32.  Nodes not understanding this option will remove the
         option.



      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

      Identification

         Copied from the Identification field of the Acknowledgement
         Request option of the packet being acknowledged.



Johnson, et al.               Experimental                     [Page 47]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      ACK Source Address

         The address of the node originating the acknowledgement.

      ACK Destination Address

         The address of the node to which the acknowledgement is to be
         delivered.

   An Acknowledgement option MAY appear one or more times within a DSR
   Options header.

6.7.  DSR Source Route Option



   The DSR Source Route option in a DSR Options header is encoded as
   follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  |  Opt Data Len |F|L|Reservd|Salvage| Segs Left |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[1]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[2]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                              ...                              |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                           Address[n]                          |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Option Type

         96.  Nodes not understanding this option will drop the packet.



      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.  For the
         format of the DSR Source Route option defined here, this field
         MUST be set to the value (n * 4) + 2, where n is the number of
         addresses present in the Address[i] fields.

      First Hop External (F)

         Set to indicate that the first hop indicated by the DSR Source
         Route option is actually an arbitrary path in a network
         external to the DSR network; the exact route outside the DSR



Johnson, et al.               Experimental                     [Page 48]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


         network is not represented in the DSR Source Route option.
         Nodes caching this hop in their Route Cache MUST flag the
         cached hop with the External flag.  Such hops MUST NOT be
         returned in a Route Reply generated from this Route Cache
         entry, and selection of routes from the Route Cache to route a
         packet being sent SHOULD prefer routes that contain no hops
         flagged as External.

      Last Hop External (L)

         Set to indicate that the last hop indicated by the DSR Source
         Route option is actually an arbitrary path in a network
         external to the DSR network; the exact route outside the DSR
         network is not represented in the DSR Source Route option.
         Nodes caching this hop in their Route Cache MUST flag the
         cached hop with the External flag.  Such hops MUST NOT be
         returned in a Route Reply generated from this Route Cache
         entry, and selection of routes from the Route Cache to route a
         packet being sent SHOULD prefer routes that contain no hops
         flagged as External.

      Reserved

         MUST be sent as 0 and ignored on reception.

      Salvage

         A 4-bit unsigned integer.  Count of number of times that this
         packet has been salvaged as a part of DSR routing (Section
         3.4.1).

      Segments Left (Segs Left)

         Number of route segments remaining, i.e., number of explicitly
         listed intermediate nodes still to be visited before reaching
         the final destination.

      Address[1..n]

         The sequence of addresses of the source route.  In routing and
         forwarding the packet, the source route is processed as
         described in Sections 8.1.3 and 8.1.5.  The number of addresses
         present in the Address[1..n] field is indicated by the Opt Data
         Len field in the option (n = (Opt Data Len - 2) / 4).

   When forwarding a packet along a DSR source route using a DSR Source
   Route option in the packet's DSR Options header, the Destination
   Address field in the packet's IP header is always set to the address



Johnson, et al.               Experimental                     [Page 49]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   of the packet's ultimate destination.  A node receiving a packet
   containing a DSR Options header with a DSR Source Route option MUST
   examine the indicated source route to determine if it is the intended
   next-hop node for the packet and how to forward the packet, as
   defined in Sections 8.1.4 and 8.1.5.

6.8.  Pad1 Option



   The Pad1 option in a DSR Options header is encoded as follows:

   +-+-+-+-+-+-+-+-+
   |  Option Type  |
   +-+-+-+-+-+-+-+-+

      Option Type

         224.  Nodes not understanding this option will drop the packet
         and return a Route Error.



   A Pad1 option MAY be included in the Options field of a DSR Options
   header in order to align subsequent DSR options, but such alignment
   is not required and MUST NOT be expected by a node receiving a packet
   containing a DSR Options header.

   If any headers follow the DSR Options header in a packet, the total
   length of a DSR Options header, indicated by the Payload Length field
   in the DSR Options header MUST be a multiple of 4 octets.  In this
   case, when building a DSR Options header in a packet, sufficient Pad1
   or PadN options MUST be included in the Options field of the DSR
   Options header to make the total length a multiple of 4 octets.

   If more than one consecutive octet of padding is being inserted in
   the Options field of a DSR Options header, the PadN option described
   next, SHOULD be used, rather than multiple Pad1 options.

   Note that the format of the Pad1 option is a special case; it does
   not have an Opt Data Len or Option Data field.

6.9.  PadN Option



   The PadN option in a DSR Options header is encoded as follows:

   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - -
   |  Option Type  |  Opt Data Len |   Option Data
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+- - - - - - - - -






Johnson, et al.               Experimental                     [Page 50]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Option Type

         0.  Nodes not understanding this option will ignore this
         option.

      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.  The size of
         the Option Data field.

      Option Data

         A number of zero-valued octets equal to the Opt Data Len.

   A PadN option MAY be included in the Options field of a DSR Options
   header in order to align subsequent DSR options, but such alignment
   is not required and MUST NOT be expected by a node receiving a packet
   containing a DSR Options header.

   If any headers follow the DSR Options header in a packet, the total
   length of a DSR Options header, indicated by the Payload Length field
   in the DSR Options header, MUST be a multiple of 4 octets.  In this
   case, when building a DSR Options header in a packet, sufficient Pad1
   or PadN options MUST be included in the Options field of the DSR
   Options header to make the total length a multiple of 4 octets.

7.  Additional Header Formats and Options for Flow State Extension



   The optional DSR flow state extension requires a new header type, the
   DSR Flow State header.

   In addition, the DSR flow state extension adds the following options
   for the DSR Options header defined in Section 6:

   -  Timeout option (Section 7.2.1)

   -  Destination and Flow ID option (Section 7.2.2)

   Two new Error Type values are also defined for use in the Route Error
   option in a DSR Options header:

   -  UNKNOWN_FLOW

   -  DEFAULT_FLOW_UNKNOWN

   Finally, an extension to the Acknowledgement Request option in a DSR
   Options header is also defined:



Johnson, et al.               Experimental                     [Page 51]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Previous Hop Address

   This section defines each of these new header, option, or extension
   formats.

7.1.  DSR Flow State Header



   The DSR Flow State header is a small 4-byte header optionally used to
   carry the flow ID and hop count for a packet being sent along a DSR
   flow.  It is distinguished from the fixed DSR Options header (Section
   6.1) in that the Flow State Header (F) bit is set in the DSR Flow
   State header and is clear in the fixed DSR Options header.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Next Header  |F|  Hop Count  |        Flow Identifier        |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Next Header

         8-bit selector.  Identifies the type of header immediately
         following the DSR Flow State header.  Uses the same values as
         the IPv4 Protocol field [RFC1700].

      Flow State Header (F)

         Flag bit.  MUST be set to 1.  This bit is set in a DSR Flow
         State header and clear in a DSR Options header (Section 6.1).

      Hop Count

         7-bit unsigned integer.  The number of hops through which this
         packet has been forwarded.

      Flow Identification

         The flow ID for this flow, as described in Section 3.5.1.

7.2.  New Options and Extensions in DSR Options Header



7.2.1.  Timeout Option



   The Timeout option is defined for use in a DSR Options header to
   indicate the amount of time before the expiration of the flow ID
   along which the packet is being sent.





Johnson, et al.               Experimental                     [Page 52]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  | Opt Data Len  |            Timeout            |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Option Type

         128.  Nodes not understanding this option will ignore the
         option and return a Route Error.



      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

         When no extensions are present, the Opt Data Len of a Timeout
         option is 2.  Further extensions to DSR may include additional
         data in a Timeout option.  The presence of such extensions is
         indicated by an Opt Data Len greater than 2.  Currently, no
         such extensions have been defined.

      Timeout

         The number of seconds for which this flow remains valid.

   The Timeout option MUST NOT appear more than once within a DSR
   Options header.

7.2.2.  Destination and Flow ID Option



   The Destination and Flow ID option is defined for use in a DSR
   Options header to send a packet to an intermediate host along one
   flow, for eventual forwarding to the final destination along a
   different flow.  This option enables the aggregation of the state of
   multiple flows.

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  | Opt Data Len  |      New Flow Identifier      |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                   New IP Destination Address                  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+







Johnson, et al.               Experimental                     [Page 53]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Option Type

         129.  Nodes not understanding this option will ignore the
         option and return a Route Error.



      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

         When no extensions are present, the Opt Data Len of a
         Destination and Flow ID option is 6.  Further extensions to DSR
         may include additional data in a Destination and Flow ID
         option.  The presence of such extensions is indicated by an Opt
         Data Len greater than 6.  Currently, no such extensions have
         been defined.

      New Flow Identifier

         Indicates the next identifier to store in the Flow ID field of
         the DSR Options header.

      New IP Destination Address

         Indicates the next address to store in the Destination Address
         field of the IP header.

   The Destination and Flow ID option MAY appear one or more times
   within a DSR Options header.

7.3.  New Error Types for Route Error Option



7.3.1.  Unknown Flow Type-Specific Information



   A new Error Type value of 129 (UNKNOWN_FLOW) is defined for use in a
   Route Error option in a DSR Options header.  The Type-Specific
   Information for errors of this type is encoded as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                Original IP Destination Address                |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |            Flow ID            |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+






Johnson, et al.               Experimental                     [Page 54]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Original IP Destination Address

         The IP Destination Address of the packet that caused the error.

      Flow ID

         The Flow ID contained in the DSR Flow ID option that caused the
         error.

7.3.2.  Default Flow Unknown Type-Specific Information



   A new Error Type value of 130 (DEFAULT_FLOW_UNKNOWN) is defined
   for use in a Route Error option in a DSR Options header.  The
   Type-Specific Information for errors of this type is encoded as
   follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |               Original IP Destination Address                 |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Original IP Destination Address

         The IP Destination Address of the packet that caused the error.

7.4.  New Acknowledgement Request Option Extension



7.4.1.  Previous Hop Address Extension



   When the Opt Data Len field of an Acknowledgement Request option
   in a DSR Options header is greater than or equal to 6, the
   ACK Request Source Address field is present.  The option is then
   formatted as follows:

    0                   1                   2                   3
    0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |  Option Type  | Opt Data Len  |       Packet Identifier       |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
   |                   ACK Request Source Address                  |
   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

      Option Type

         160.  Nodes not understanding this option will remove the
         option and return a Route Error.




Johnson, et al.               Experimental                     [Page 55]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Opt Data Len

         8-bit unsigned integer.  Length of the option, in octets,
         excluding the Option Type and Opt Data Len fields.

         When no extensions are presents, the Opt Data Len of an
         Acknowledgement Request option is 2.  Further extensions to DSR
         may include additional data in an Acknowledgement Request
         option.  The presence of such extensions is indicated by an Opt
         Data Len greater than 2.

         Currently, one such extension has been defined.  If the Opt
         Data Len is at least 6, then an ACK Request Source Address is
         present.

      Packet Identifier

         The Packet Identifier field is set to a unique number and is
         copied into the Identification field of the DSR Acknowledgement
         option when returned by the node receiving the packet over this
         hop.

      ACK Request Source Address

         The address of the node requesting the DSR Acknowledgement.

8.  Detailed Operation



8.1.  General Packet Processing



8.1.1.  Originating a Packet



   When originating any packet, a node using DSR routing MUST perform
   the following sequence of steps:

   -  Search the node's Route Cache for a route to the address given in
      the IP Destination Address field in the packet's header.

   -  If no such route is found in the Route Cache, then perform Route
      Discovery for the Destination Address, as described in Section
      8.2.  Initiating a Route Discovery for this target node address
      results in the node adding a Route Request option in a DSR Options
      header in this existing packet, or saving this existing packet to
      its Send Buffer and initiating the Route Discovery by sending a
      separate packet containing such a Route Request option.  If the
      node chooses to initiate the Route Discovery by adding the Route
      Request option to this existing packet, it will replace the IP
      Destination Address field with the IP "limited broadcast" address



Johnson, et al.               Experimental                     [Page 56]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      (255.255.255.255) [RFC1122], copying the original IP Destination
      Address to the Target Address field of the new Route Request
      option added to the packet, as described in Section 8.2.1.

   -  If the packet now does not contain a Route Request option, then
      this node must have a route to the Destination Address of the
      packet; if the node has more than one route to this Destination
      Address, the node selects one to use for this packet.  If the
      length of this route is greater than 1 hop, or if the node
      determines to request a DSR network-layer acknowledgement from the
      first-hop node in that route, then insert a DSR Options header
      into the packet, as described in Section 8.1.2, and insert a DSR
      Source Route option, as described in Section 8.1.3.  The source
      route in the packet is initialized from the selected route to the
      Destination Address of the packet.

   -  Transmit the packet to the first-hop node address given in
      selected source route, using Route Maintenance to determine the
      reachability of the next hop, as described in Section 8.3.

8.1.2.  Adding a DSR Options Header to a Packet



   A node originating a packet adds a DSR Options header to the packet,
   if necessary, to carry information needed by the routing protocol.  A
   packet MUST NOT contain more than one DSR Options header.  A DSR
   Options header is added to a packet by performing the following
   sequence of steps (these steps assume that the packet contains no
   other headers that MUST be located in the packet before the DSR
   Options header):

   -  Insert a DSR Options header after the IP header but before any
      other header that may be present.

   -  Set the Next Header field of the DSR Options header to the
      Protocol number field of the packet's IP header.

   -  Set the Protocol field of the packet's IP header to the protocol
      number assigned for DSR (48).

8.1.3.  Adding a DSR Source Route Option to a Packet



   A node originating a packet adds a DSR Source Route option to the
   packet, if necessary, in order to carry the source route from this
   originating node to the final destination address of the packet.
   Specifically, the node adding the DSR Source Route option constructs
   the DSR Source Route option and modifies the IP packet according to
   the following sequence of steps:




Johnson, et al.               Experimental                     [Page 57]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  The node creates a DSR Source Route option, as described in
      Section 6.7, and appends it to the DSR Options header in the
      packet.  (A DSR Options header is added, as described in Section
      8.1.2, if not already present.)

   -  The number of Address[i] fields to include in the DSR Source Route
      option (n) is the number of intermediate nodes in the source route
      for the packet (i.e., excluding the address of the originating
      node and the final destination address of the packet).  The
      Segments Left field in the DSR Source Route option is initialized
      equal to n.

   -  The addresses within the source route for the packet are copied
      into sequential Address[i] fields in the DSR Source Route option,
      for i = 1, 2, ..., n.

   -  The First Hop External (F) bit in the DSR Source Route option is
      copied from the External bit flagging the first hop in the source
      route for the packet, as indicated in the Route Cache.

   -  The Last Hop External (L) bit in the DSR Source Route option is
      copied from the External bit flagging the last hop in the source
      route for the packet, as indicated in the Route Cache.

   -  The Salvage field in the DSR Source Route option is initialized to
      0.

8.1.4.  Processing a Received Packet



   When a node receives any packet (whether for forwarding, overheard,
   or the final destination of the packet), if that packet contains a
   DSR Options header, then that node MUST process any options contained
   in that DSR Options header, in the order contained there.
   Specifically:

   -  If the DSR Options header contains a Route Request option, the
      node SHOULD extract the source route from the Route Request and
      add this routing information to its Route Cache, subject to the
      conditions identified in Section 3.3.1.  The routing information
      from the Route Request is the sequence of hop addresses

         initiator, Address[1], Address[2], ..., Address[n]

      where initiator is the value of the Source Address field in the IP
      header of the packet carrying the Route Request (the address of
      the initiator of the Route Discovery), and each Address[i] is a
      node through which this Route Request has passed, in turn, during




Johnson, et al.               Experimental                     [Page 58]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      this Route Discovery.  The value n, here, is the number of
      addresses recorded in the Route Request option, or
      (Opt Data Len - 6) / 4.

      After possibly updating the node's Route Cache in response to the
      routing information in the Route Request option, the node MUST
      then process the Route Request option as described in Section
      8.2.2.

   -  If the DSR Options header contains a Route Reply option, the node
      SHOULD extract the source route from the Route Reply and add this
      routing information to its Route Cache, subject to the conditions
      identified in Section 3.3.1.  The source route from the Route
      Reply is the sequence of hop addresses

         initiator, Address[1], Address[2], ..., Address[n]

      where initiator is the value of the Destination Address field in
      the IP header of the packet carrying the Route Reply (the address
      of the initiator of the Route Discovery), and each Address[i] is a
      node through which the source route passes, in turn, on the route
      to the target of the Route Discovery.  Address[n] is the address
      of the target.  If the Last Hop External (L) bit is set in the
      Route Reply, the node MUST flag the last hop from the Route Reply
      (the link from Address[n-1] to Address[n]) in its Route Cache as
      External.  The value n here is the number of addresses in the
      source route being returned in the Route Reply option, or
      (Opt Data Len - 1) / 4.

      After possibly updating the node's Route Cache in response to the
      routing information in the Route Reply option, then if the
      packet's IP Destination Address matches one of this node's IP
      addresses, the node MUST then process the Route Reply option as
      described in Section 8.2.6.

   -  If the DSR Options header contains a Route Error option, the node
      MUST process the Route Error option as described in Section 8.3.5.

   -  If the DSR Options header contains an Acknowledgement Request
      option, the node MUST process the Acknowledgement Request option
      as described in Section 8.3.3.

   -  If the DSR Options header contains an Acknowledgement option, then
      subject to the conditions identified in Section 3.3.1, the node
      SHOULD add to its Route Cache the single link from the node
      identified by the ACK Source Address field to the node identified
      by the ACK Destination Address field.




Johnson, et al.               Experimental                     [Page 59]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      After possibly updating the node's Route Cache in response to the
      routing information in the Acknowledgement option, the node MUST
      then process the Acknowledgement option as described in Section
      8.3.3.

   -  If the DSR Options header contains a DSR Source Route option, the
      node SHOULD extract the source route from the DSR Source Route
      option and add this routing information to its Route Cache,
      subject to the conditions identified in Section 3.3.1.  If the
      value of the Salvage field in the DSR Source Route option is zero,
      then the routing information from the DSR Source Route is the
      sequence of hop addresses

         source, Address[1], Address[2], ..., Address[n], destination

      Otherwise (i.e., if Salvage is nonzero), the routing information
      from the DSR Source Route is the sequence of hop addresses

         Address[1], Address[2], ..., Address[n], destination

      where source is the value of the Source Address field in the IP
      header of the packet carrying the DSR Source Route option (the
      original sender of the packet), each Address[i] is the value in
      the Address[i] field in the DSR Source Route option, and
      destination is the value of the Destination Address field in the
      packet's IP header (the last-hop address of the source route).
      The value n here is the number of addresses in source route in the
      DSR Source Route option, or (Opt Data Len - 2) / 4.

      After possibly updating the node's Route Cache in response to the
      routing information in the DSR Source Route option, the node MUST
      then process the DSR Source Route option as described in Section
      8.1.5.

   -  Any Pad1 or PadN options in the DSR Options header are ignored.

   -  Finally, if the Destination Address in the packet's IP header
      matches one of this receiving node's own IP address(es), remove
      the DSR Options header and all the included DSR options in the
      header, and pass the rest of the packet to the network layer.

8.1.5.  Processing a Received DSR Source Route Option



   When a node receives a packet containing a DSR Source Route option
   (whether for forwarding, overheard, or the final destination of the
   packet), that node SHOULD examine the packet to determine if the
   receipt of that packet indicates an opportunity for automatic route
   shortening, as described in Section 3.4.3.  Specifically, if this



Johnson, et al.               Experimental                     [Page 60]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   node is not the intended next-hop destination for the packet but is
   named in the later unexpended portion of the source route in the
   packet's DSR Source Route option, then this packet indicates an
   opportunity for automatic route shortening:  the intermediate nodes
   after the node from which this node overheard the packet and before
   this node itself are no longer necessary in the source route.  In
   this case, this node SHOULD perform the following sequence of steps
   as part of automatic route shortening:

   -  The node searches its Gratuitous Route Reply Table for an entry
      describing a gratuitous Route Reply earlier sent by this node, for
      which the original sender (of the packet triggering the gratuitous
      Route Reply) and the transmitting node (from which this node
      overheard that packet in order to trigger the gratuitous Route
      Reply) both match the respective node addresses for this new
      received packet.  If such an entry is found in the node's
      Gratuitous Route Reply Table, the node SHOULD NOT perform
      automatic route shortening in response to this receipt of this
      packet.

   -  Otherwise, the node creates an entry for this overheard packet in
      its Gratuitous Route Reply Table.  The timeout value for this new
      entry SHOULD be initialized to the value GratReplyHoldoff.  After
      this timeout has expired, the node SHOULD delete this entry from
      its Gratuitous Route Reply Table.

   -  After creating the new Gratuitous Route Reply Table entry above,
      the node originates a gratuitous Route Reply to the IP Source
      Address of this overheard packet, as described in Section 3.4.3.

      If the MAC protocol in use in the network is not capable of
      transmitting unicast packets over unidirectional links, as
      discussed in Section 3.3.1, then in originating this Route Reply,
      the node MUST use a source route for routing the Route Reply
      packet that is obtained by reversing the sequence of hops over
      which the packet triggering the gratuitous Route Reply was routed
      in reaching and being overheard by this node.  This reversing of
      the route uses the gratuitous Route Reply to test this sequence of
      hops for bidirectionality, preventing the gratuitous Route Reply
      from being received by the initiator of the Route Discovery unless
      each of the hops over which the gratuitous Route Reply is returned
      is bidirectional.

   -  Discard the overheard packet, since the packet has been received
      before its normal traversal of the packet's source route would
      have caused it to reach this receiving node.  Another copy of the
      packet will normally arrive at this node as indicated in the




Johnson, et al.               Experimental                     [Page 61]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      packet's source route; discarding this initial copy of the packet,
      which triggered the gratuitous Route Reply, will prevent the
      duplication of this packet that would otherwise occur.

   If the packet is not discarded as part of automatic route shortening
   above, then the node MUST process the Source Route option according
   to the following sequence of steps:

   -  If the value of the Segments Left field in the DSR Source Route
      option equals 0, then remove the DSR Source Route option from the
      DSR Options header.

   -  Else, let n equal (Opt Data Len - 2) / 4.  This is the number of
      addresses in the DSR Source Route option.

   -  If the value of the Segments Left field is greater than n, then
      send an ICMP Parameter Problem, Code 0, message [RFC792] to the IP
      Source Address, pointing to the Segments Left field, and discard
      the packet.  Do not process the DSR Source Route option further.

   -  Else, decrement the value of the Segments Left field by 1.  Let i
      equal n minus Segments Left.  This is the index of the next
      address to be visited in the Address vector.

   -  If Address[i] or the IP Destination Address is a multicast
      address, then discard the packet.  Do not process the DSR Source
      Route option further.

   -  If this node has more than one network interface and if Address[i]
      is the address of one this node's network interfaces, then this
      indicates a change in the network interface to use in forwarding
      the packet, as described in Section 8.4.  In this case, decrement
      the value of the Segments Left field by 1 to skip over this
      address (that indicated the change of network interface) and go to
      the first step above (checking the value of the Segments Left
      field) to continue processing this Source Route option; in further
      processing of this Source Route option, the indicated new network
      interface MUST be used in forwarding the packet.

   -  If the MTU of the link over which this node would transmit the
      packet to forward it to the node Address[i] is less than the size
      of the packet, the node MUST either discard the packet and send an
      ICMP Packet Too Big message to the packet's Source Address
      [RFC792] or fragment it as specified in Section 8.5.

   -  Forward the packet to the IP address specified in the Address[i]
      field of the IP header, following normal IP forwarding procedures,
      including checking and decrementing the Time-to-Live (TTL) field



Johnson, et al.               Experimental                     [Page 62]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      in the packet's IP header [RFC791, RFC1122].  In this forwarding
      of the packet, the next-hop node (identified by Address[i]) MUST
      be treated as a direct neighbor node:  the transmission to that
      next node MUST be done in a single IP forwarding hop, without
      Route Discovery and without searching the Route Cache.

   -  In forwarding the packet, perform Route Maintenance for the next
      hop of the packet, by verifying that the next-hop node is
      reachable, as described in Section 8.3.

   Multicast addresses MUST NOT appear in a DSR Source Route option or
   in the IP Destination Address field of a packet carrying a DSR Source
   Route option in a DSR Options header.

8.1.6.  Handling an Unknown DSR Option



   Nodes implementing DSR MUST handle all options specified in this
   document, except those options pertaining to the optional flow state
   extension (Section 7).  However, further extensions to DSR may
   include other option types that may not be understood by
   implementations conforming to this version of the DSR specification.
   In DSR, Option Type codes encode required behavior for nodes not
   implementing that type of option.  These behaviors are included in
   the most significant 3 bits of the Option Type.

   If the most significant bit of the Option Type is set (that is,
   Option Type & 0x80 is nonzero), and this packet does not contain a
   Route Request option, a node SHOULD return a Route Error to the IP
   Source Address, following the steps described in Section 8.3.4,
   except that the Error Type MUST be set to OPTION_NOT_SUPPORTED and
   the Unsupported Opt field MUST be set to the Option Type triggering
   the Route Error.

   Whether or not a Route Error is sent in response to this DSR option,
   as described above, the node also MUST examine the next 2 most
   significant bits (that is, Option Type & 0x60):

   -  When these 2 bits are 00 (that is, Option Type & 0x60 == 0), a
      node not implementing processing for that Option Type MUST use the
      Opt Data Len field to skip over the option and continue
      processing.

   -  When these 2 bits are 01 (that is, Option Type & 0x60 == 0x20), a
      node not implementing processing for that Option Type MUST use the
      Opt Data Len field to remove the option from the packet and
      continue processing as if the option had not been included in the
      received packet.




Johnson, et al.               Experimental                     [Page 63]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  When these 2 bits are 10 (that is, Option Type & 0x60 == 0x40), a
      node not implementing processing for that Option Type MUST set the
      most significant bit following the Opt Data Len field.  In
      addition, the node MUST then ignore and skip over the contents of
      the option using the Opt Data Len field and MUST continue
      processing the packet.

   -  Finally, when these 2 bits are 11 (that is,
      Option Type & 0x60 == 0x60), a node not implementing processing
      for that Option Type MUST drop the packet.

8.2.  Route Discovery Processing



   Route Discovery is the mechanism by which a node S wishing to send a
   packet to a destination node D obtains a source route to D.  Route
   Discovery SHOULD be used only when S attempts to send a packet to D
   and does not already know a route to D.  The node initiating a Route
   Discovery is known as the "initiator" of the Route Discovery, and the
   destination node for which the Route Discovery is initiated is known
   as the "target" of the Route Discovery.

   Route Discovery operates entirely on demand; a node initiates Route
   Discovery based on its own origination of new packets for some
   destination address to which it does not currently know a route.
   Route Discovery does not depend on any periodic or background
   exchange of routing information or neighbor node detection at any
   layer in the network protocol stack at any node.

   The Route Discovery procedure utilizes two types of messages, a Route
   Request (Section 6.2) and a Route Reply (Section 6.3), to actively
   search the ad hoc network for a route to the desired target
   destination.  These DSR messages MAY be carried in any type of IP
   packet, through use of the DSR Options header as described in Section
   6.

   Except as discussed in Section 8.3.5, a Route Discovery for a
   destination address SHOULD NOT be initiated unless the initiating
   node has a packet in its Send Buffer requiring delivery to that
   destination.  A Route Discovery for a given target node MUST NOT be
   initiated unless permitted by the rate-limiting information contained
   in the Route Request Table.  After each Route Discovery attempt, the
   interval between successive Route Discoveries for this target SHOULD
   be doubled, up to a maximum of MaxRequestPeriod, until a valid Route
   Reply is received for this target.







Johnson, et al.               Experimental                     [Page 64]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


8.2.1.  Originating a Route Request



   A node initiating a Route Discovery for some target creates and
   initializes a Route Request option in a DSR Options header in some IP
   packet.  This MAY be a separate IP packet, used only to carry this
   Route Request option, or the node MAY include the Route Request
   option in some existing packet that it needs to send to the target
   node (e.g., the IP packet originated by this node that caused the
   node to attempt Route Discovery for the destination address of the
   packet).  The Route Request option MUST be included in a DSR Options
   header in the packet.  To initialize the Route Request option, the
   node performs the following sequence of steps:

   -  The Option Type in the option MUST be set to the value 2.

   -  The Opt Data Len field in the option MUST be set to the value 6.
      The total size of the Route Request option, when initiated, is 8
      octets; the Opt Data Len field excludes the size of the Option
      Type and Opt Data Len fields themselves.

   -  The Identification field in the option MUST be set to a new value,
      different from that used for other Route Requests recently
      initiated by this node for this same target address.  For example,
      each node MAY maintain a single counter value for generating a new
      Identification value for each Route Request it initiates.

   -  The Target Address field in the option MUST be set to the IP
      address that is the target of this Route Discovery.

   The Source Address in the IP header of this packet MUST be the node's
   own IP address.  The Destination Address in the IP header of this
   packet MUST be the IP "limited broadcast" address (255.255.255.255).

   A node MUST maintain, in its Route Request Table, information about
   Route Requests that it initiates.  When initiating a new Route
   Request, the node MUST use the information recorded in the Route
   Request Table entry for the target of that Route Request, and it MUST
   update that information in the table entry for use in the next Route
   Request initiated for this target.  In particular:

   -  The Route Request Table entry for a target node records the Time-
      to-Live (TTL) field used in the IP header of the Route Request for
      the last Route Discovery initiated by this node for that target
      node.  This value allows the node to implement a variety of
      algorithms for controlling the spread of its Route Request on each
      Route Discovery initiated for a target.  As examples, two possible
      algorithms for this use of the TTL field are described in Section
      3.3.3.



Johnson, et al.               Experimental                     [Page 65]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  The Route Request Table entry for a target node records the number
      of consecutive Route Requests initiated for this target since
      receiving a valid Route Reply giving a route to that target node,
      and the remaining amount of time before which this node MAY next
      attempt at a Route Discovery for that target node.

      A node MUST use these values to implement a back-off algorithm to
      limit the rate at which this node initiates new Route Discoveries
      for the same target address.  In particular, until a valid Route
      Reply is received for this target node address, the timeout
      between consecutive Route Discovery initiations for this target
      node with the same hop limit SHOULD increase by doubling the
      timeout value on each new initiation.

   The behavior of a node processing a packet containing DSR Options
   header with both a DSR Source Route option and a Route Request option
   is unspecified.  Packets SHOULD NOT contain both a DSR Source Route
   option and a Route Request option.

   Packets containing a Route Request option SHOULD NOT include an
   Acknowledgement Request option, SHOULD NOT expect link-layer
   acknowledgement or passive acknowledgement, and SHOULD NOT be
   retransmitted.  The retransmission of packets containing a Route
   Request option is controlled solely by the logic described in this
   section.

8.2.2.  Processing a Received Route Request Option



   When a node receives a packet containing a Route Request option, that
   node MUST process the option according to the following sequence of
   steps:

   -  If the Target Address field in the Route Request matches this
      node's own IP address, then the node SHOULD return a Route Reply
      to the initiator of this Route Request (the Source Address in the
      IP header of the packet), as described in Section 8.2.4.  The
      source route for this Reply is the sequence of hop addresses

         initiator, Address[1], Address[2], ..., Address[n], target

      where initiator is the address of the initiator of this Route
      Request, each Address[i] is an address from the Route Request, and
      target is the target of the Route Request (the Target Address
      field in the Route Request).  The value n here is the number of
      addresses recorded in the Route Request, or
      (Opt Data Len - 6) / 4.





Johnson, et al.               Experimental                     [Page 66]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      The node then MUST replace the Destination Address field in the
      Route Request packet's IP header with the value in the Target
      Address field in the Route Request option, and continue processing
      the rest of the Route Request packet normally.  The node MUST NOT
      process the Route Request option further and MUST NOT retransmit
      the Route Request to propagate it to other nodes as part of the
      Route Discovery.

   -  Else, the node MUST examine the route recorded in the Route
      Request option (the IP Source Address field and the sequence of
      Address[i] fields) to determine if this node's own IP address
      already appears in this list of addresses.  If so, the node MUST
      discard the entire packet carrying the Route Request option.

   -  Else, if the Route Request was received through a network
      interface that requires physically bidirectional links for unicast
      transmission, the node MUST check if the Route Request was last
      forwarded by a node on its blacklist (Section 4.6).  If such an
      entry is found in the blacklist, and the state of the
      unidirectional link is "probable", then the Request MUST be
      silently discarded.

   -  Else, if the Route Request was received through a network
      interface that requires physically bidirectional links for unicast
      transmission, the node MUST check if the Route Request was last
      forwarded by a node on its blacklist.  If such an entry is found
      in the blacklist, and the state of the unidirectional link is
      "questionable", then the node MUST create and unicast a Route
      Request packet to that previous node, setting the IP Time-To-Live
      (TTL) to 1 to prevent the Request from being propagated.  If the
      node receives a Route Reply in response to the new Request, it
      MUST remove the blacklist entry for that node, and SHOULD continue
      processing.  If the node does not receive a Route Reply within
      some reasonable amount of time, the node MUST silently discard the
      Route Request packet.

   -  Else, the node MUST search its Route Request Table for an entry
      for the initiator of this Route Request (the IP Source Address
      field).  If such an entry is found in the table, the node MUST
      search the cache of Identification values of recently received
      Route Requests in that table entry, to determine if an entry is
      present in the cache matching the Identification value and target
      node address in this Route Request.  If such an (Identification,
      target address) entry is found in this cache in this entry in the
      Route Request Table, then the node MUST discard the entire packet
      carrying the Route Request option.





Johnson, et al.               Experimental                     [Page 67]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Else, this node SHOULD further process the Route Request according
      to the following sequence of steps:

      o  Add an entry for this Route Request in its cache of
         (Identification, target address) values of recently received
         Route Requests.

      o  Conceptually create a copy of this entire packet and perform
         the following steps on the copy of the packet.

      o  Append this node's own IP address to the list of Address[i]
         values in the Route Request and increase the value of the Opt
         Data Len field in the Route Request by 4 (the size of an IP
         address).  However, if the node has multiple network
         interfaces, this step MUST be modified by the special
         processing specified in Section 8.4.

      o  This node SHOULD search its own Route Cache for a route (from
         itself, as if it were the source of a packet) to the target of
         this Route Request.  If such a route is found in its Route
         Cache, then this node SHOULD follow the procedure outlined in
         Section 8.2.3 to return a "cached Route Reply" to the initiator
         of this Route Request, if permitted by the restrictions
         specified there.

      o  If the node does not return a cached Route Reply, then this
         node SHOULD transmit this copy of the packet as a link-layer
         broadcast, with a short jitter delay before the broadcast is
         sent.  The jitter period SHOULD be chosen as a random period,
         uniformly distributed between 0 and BroadcastJitter.

8.2.3.  Generating a Route Reply Using the Route Cache



   As described in Section 3.3.2, it is possible for a node processing a
   received Route Request to avoid propagating the Route Request further
   toward the target of the Request, if this node has in its Route Cache
   a route from itself to this target.  Such a Route Reply generated by
   a node from its own cached route to the target of a Route Request is
   called a "cached Route Reply", and this mechanism can greatly reduce
   the overall overhead of Route Discovery on the network by reducing
   the flood of Route Requests.  The general processing of a received
   Route Request is described in Section 8.2.2; this section specifies
   the additional requirements that MUST be met before a cached Route
   Reply may be generated and returned and specifies the procedure for
   returning such a cached Route Reply.






Johnson, et al.               Experimental                     [Page 68]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   While processing a received Route Request, for a node to possibly
   return a cached Route Reply, it MUST have in its Route Cache a route
   from itself to the target of this Route Request.  However, before
   generating a cached Route Reply for this Route Request, the node MUST
   verify that there are no duplicate addresses listed in the route
   accumulated in the Route Request together with the route from this
   node's Route Cache.  Specifically, there MUST be no duplicates among
   the following addresses:

   -  The IP Source Address of the packet containing the Route Request,

   -  The Address[i] fields in the Route Request, and

   -  The nodes listed in the route obtained from this node's Route
      Cache, excluding the address of this node itself (this node itself
      is the common point between the route accumulated in the Route
      Request and the route obtained from the Route Cache).

   If any duplicates exist among these addresses, then the node MUST NOT
   send a cached Route Reply using this route from the Route Cache (it
   is possible that this node has another route in its Route Cache for
   which the above restriction on duplicate addresses is met, allowing
   the node to send a cached Route Reply based on that cached route,
   instead).  The node SHOULD continue to process the Route Request as
   described in Section 8.2.2 if it does not send a cached Route Reply.

   If the Route Request and the route from the Route Cache meet the
   restriction above, then the node SHOULD construct and return a cached
   Route Reply as follows:

   -  The source route for this Route Reply is the sequence of hop
      addresses

         initiator, Address[1], Address[2], ..., Address[n], c-route

      where initiator is the address of the initiator of this Route
      Request, each Address[i] is an address from the Route Request, and
      c-route is the sequence of hop addresses in the source route to
      this target node, obtained from the node's Route Cache.  In
      appending this cached route to the source route for the reply, the
      address of this node itself MUST be excluded, since it is already
      listed as Address[n].

   -  Send a Route Reply to the initiator of the Route Request, using
      the procedure defined in Section 8.2.4.  The initiator of the
      Route Request is indicated in the Source Address field in the
      packet's IP header.




Johnson, et al.               Experimental                     [Page 69]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   Before sending the cached Route Reply, however, the node MAY delay
   the Reply in order to help prevent a possible Route Reply "storm", as
   described in Section 8.2.5.

   If the node returns a cached Route Reply as described above, then the
   node MUST NOT propagate the Route Request further (i.e., the node
   MUST NOT rebroadcast the Route Request).  In this case, instead, if
   the packet contains no other DSR options and contains no payload
   after the DSR Options header (e.g., the Route Request is not
   piggybacked on a TCP or UDP packet), then the node SHOULD simply
   discard the packet.  Otherwise (if the packet contains other DSR
   options or contains any payload after the DSR Options header), the
   node SHOULD forward the packet along the cached route to the target
   of the Route Request.  Specifically, if the node does so, it MUST use
   the following steps:

   -  Copy the Target Address from the Route Request option in the DSR
      Options header to the Destination Address field in the packet's IP
      header.

   -  Remove the Route Request option from the DSR Options header in the
      packet, and add a DSR Source Route option to the packet's DSR
      Options header.

   -  In the DSR Source Route option, set the Address[i] fields to
      represent the source route found in this node's Route Cache to the
      original target of the Route Discovery (the new IP Destination
      Address of the packet).  Specifically, the node copies the hop
      addresses of the source route into sequential Address[i] fields in
      the DSR Source Route option, for i = 1, 2, ..., n.  Address[1],
      here, is the address of this node itself (the first address in the
      source route found from this node to the original target of the
      Route Discovery).  The value n, here, is the number of hop
      addresses in this source route, excluding the destination of the
      packet (which is instead already represented in the Destination
      Address field in the packet's IP header).

   -  Initialize the Segments Left field in the DSR Source Route option
      to n as defined above.

   -  The First Hop External (F) bit in the DSR Source Route option MUST
      be set to 0.

   -  The Last Hop External (L) bit in the DSR Source Route option is
      copied from the External bit flagging the last hop in the source
      route for the packet, as indicated in the Route Cache.





Johnson, et al.               Experimental                     [Page 70]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  The Salvage field in the DSR Source Route option MUST be
      initialized to some nonzero value; the particular nonzero value
      used SHOULD be MAX_SALVAGE_COUNT.  By initializing this field to a
      nonzero value, nodes forwarding or overhearing this packet will
      not consider a link to exist between the IP Source Address of the
      packet and the Address[1] address in the DSR Source Route option
      (e.g., they will not attempt to add this to their Route Cache as a
      link).  By choosing MAX_SALVAGE_COUNT as the nonzero value to
      which the node initializes this field, nodes furthermore will not
      attempt to salvage this packet.

   -  Transmit the packet to the next-hop node on the new source route
      in the packet, using the forwarding procedure described in Section
      8.1.5.

8.2.4.  Originating a Route Reply



   A node originates a Route Reply in order to reply to a received and
   processed Route Request, according to the procedures described in
   Sections 8.2.2 and 8.2.3.  The Route Reply is returned in a Route
   Reply option (Section 6.3).  The Route Reply option MAY be returned
   to the initiator of the Route Request in a separate IP packet, used
   only to carry this Route Reply option, or it MAY be included in any
   other IP packet being sent to this address.

   The Route Reply option MUST be included in a DSR Options header in
   the packet returned to the initiator.  To initialize the Route Reply
   option, the node performs the following sequence of steps:

   -  The Option Type in the option MUST be set to the value 3.

   -  The Opt Data Len field in the option MUST be set to the value
      (n * 4) + 3, where n is the number of addresses in the source
      route being returned (excluding the Route Discovery initiator
      node's address).

   -  If this node is the target of the Route Request, the Last Hop
      External (L) bit in the option MUST be initialized to 0.

   -  The Reserved field in the option MUST be initialized to 0.

   -  The Route Request Identifier MUST be initialized to the Identifier
      field of the Route Request to which this Route Reply is sent in
      response.

   -  The sequence of hop addresses in the source route are copied into
      the Address[i] fields of the option.  Address[1] MUST be set to
      the first-hop address of the route after the initiator of the



Johnson, et al.               Experimental                     [Page 71]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      Route Discovery, Address[n] MUST be set to the last-hop address of
      the source route (the address of the target node), and each other
      Address[i] MUST be set to the next address in sequence in the
      source route being returned.

   The Destination Address field in the IP header of the packet carrying
   the Route Reply option MUST be set to the address of the initiator of
   the Route Discovery (i.e., for a Route Reply being returned in
   response to some Route Request, the IP Source Address of the Route
   Request).

   After creating and initializing the Route Reply option and the IP
   packet containing it, send the Route Reply.  In sending the Route
   Reply from this node (but not from nodes forwarding the Route Reply),
   this node SHOULD delay the Reply by a small jitter period chosen
   randomly between 0 and BroadcastJitter.

   When returning any Route Reply in the case in which the MAC protocol
   in use in the network is not capable of transmitting unicast packets
   over unidirectional links, the source route used for routing the
   Route Reply packet MUST be obtained by reversing the sequence of hops
   in the Route Request packet (the source route that is then returned
   in the Route Reply).  This restriction on returning a Route Reply
   enables the Route Reply to test this sequence of hops for
   bidirectionality, preventing the Route Reply from being received by
   the initiator of the Route Discovery unless each of the hops over
   which the Route Reply is returned (and thus each of the hops in the
   source route being returned in the Reply) is bidirectional.

   If sending a Route Reply to the initiator of the Route Request
   requires performing a Route Discovery, the Route Reply option MUST be
   piggybacked on the packet that contains the Route Request.  This
   piggybacking prevents a recursive dependency wherein the target of
   the new Route Request (which was itself the initiator of the original
   Route Request) must do another Route Request in order to return its
   Route Reply.

   If sending the Route Reply to the initiator of the Route Request does
   not require performing a Route Discovery, a node SHOULD send a
   unicast Route Reply in response to every Route Request it receives
   for which it is the target node.

8.2.5.  Preventing Route Reply Storms



   The ability for nodes to reply to a Route Request based on
   information in their Route Caches, as described in Sections 3.3.2 and
   8.2.3, could result in a possible Route Reply "storm" in some cases.
   In particular, if a node broadcasts a Route Request for a target node



Johnson, et al.               Experimental                     [Page 72]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   for which the node's neighbors have a route in their Route Caches,
   each neighbor may attempt to send a Route Reply, thereby wasting
   bandwidth and possibly increasing the number of network collisions in
   the area.

   For example, the figure below shows a situation in which nodes B, C,
   D, E, and F all receive A's Route Request for target G, and each has
   the indicated route cached for this target:

                +-----+                 +-----+
                |  D  |<               >|  C  |
                +-----+ \             / +-----+
      Cache: C - B - G   \           /  Cache: B - G
                          \ +-----+ /
                           -|  A  |-
                            +-----+\     +-----+     +-----+
                             |   |  \--->|  B  |     |  G  |
                            /     \      +-----+     +-----+
                           /       \     Cache: G
                          v         v
                    +-----+         +-----+
                    |  E  |         |  F  |
                    +-----+         +-----+
               Cache: F - B - G     Cache: B - G

   Normally, each of these nodes would attempt to reply from its own
   Route Cache, and they would thus all send their Route Replies at
   about the same time, since they all received the broadcast Route
   Request at about the same time.  Such simultaneous Route Replies from
   different nodes all receiving the Route Request may cause local
   congestion in the wireless network and may create packet collisions
   among some or all of these Replies if the MAC protocol in use does
   not provide sufficient collision avoidance for these packets.  In
   addition, it will often be the case that the different replies will
   indicate routes of different lengths, as shown in this example.

   In order to reduce these effects, if a node can put its network
   interface into promiscuous receive mode, it MAY delay sending its own
   Route Reply for a short period, while listening to see if the
   initiating node begins using a shorter route first.  Specifically,
   this node MAY delay sending its own Route Reply for a random period

      d = H * (h - 1 + r)

   where h is the length in number of network hops for the route to be
   returned in this node's Route Reply, r is a random floating point
   number between 0 and 1, and H is a small constant delay (at least
   twice the maximum wireless link propagation delay) to be introduced



Johnson, et al.               Experimental                     [Page 73]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   per hop.  This delay effectively randomizes the time at which each
   node sends its Route Reply, with all nodes sending Route Replies
   giving routes of length less than h sending their Replies before this
   node, and all nodes sending Route Replies giving routes of length
   greater than h send their Replies after this node.

   Within the delay period, this node promiscuously receives all
   packets, looking for data packets from the initiator of this Route
   Discovery destined for the target of the Route Discovery.  If such a
   data packet received by this node during the delay period uses a
   source route of length less than or equal to h, this node may infer
   that the initiator of the Route Discovery has already received a
   Route Reply giving an equally good or better route.  In this case,
   this node SHOULD cancel its delay timer and SHOULD NOT send its Route
   Reply for this Route Discovery.

8.2.6.  Processing a Received Route Reply Option



   Section 8.1.4 describes the general processing for a received packet,
   including the addition of routing information from options in the
   packet's DSR Options header to the receiving node's Route Cache.

   If the received packet contains a Route Reply, no additional special
   processing of the Route Reply option is required beyond what is
   described there.  As described in Section 4.1, anytime a node adds
   new information to its Route Cache (including the information added
   from this Route Reply option), the node SHOULD check each packet in
   its own Send Buffer (Section 4.2) to determine whether a route to
   that packet's IP Destination Address now exists in the node's Route
   Cache (including the information just added to the Cache).  If so,
   the packet SHOULD then be sent using that route and removed from the
   Send Buffer.  This general procedure handles all processing required
   for a received Route Reply option.

   When using a MAC protocol that requires bidirectional links for
   unicast transmission, a unidirectional link may be discovered by the
   propagation of the Route Request.  When the Route Reply is sent over
   the reverse path, a forwarding node may discover that the next-hop is
   unreachable.  In this case, it MUST add the next-hop address to its
   blacklist (Section 4.6).

8.3.  Route Maintenance Processing



   Route Maintenance is the mechanism by which a source node S is able
   to detect, while using a source route to some destination node D, if
   the network topology has changed such that it can no longer use its
   route to D because a link along the route no longer works.  When
   Route Maintenance indicates that a source route is broken, S can



Johnson, et al.               Experimental                     [Page 74]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   attempt to use any other route it happens to know to D or can invoke
   Route Discovery again to find a new route for subsequent packets to
   D.  Route Maintenance for this route is used only when S is actually
   sending packets to D.

   Specifically, when forwarding a packet, a node MUST attempt to
   confirm the reachability of the next-hop node, unless such
   confirmation had been received in the last MaintHoldoffTime period.
   Individual implementations MAY choose to bypass such confirmation for
   some limited number of packets, as long as those packets all fall
   within MaintHoldoffTime since the last confirmation.  If no
   confirmation is received after the retransmission of MaxMaintRexmt
   acknowledgement requests, after the initial transmission of the
   packet, and conceptually including all retransmissions provided by
   the MAC layer, the node determines that the link for this next-hop
   node of the source route is "broken".  This confirmation from the
   next-hop node for Route Maintenance can be implemented using a link-
   layer acknowledgement (Section 8.3.1), a "passive acknowledgement"
   (Section 8.3.2), or a network-layer acknowledgement (Section 8.3.3);
   the particular strategy for retransmission timing depends on the type
   of acknowledgement mechanism used.  When not using link-layer
   acknowledgements for Route Maintenance, nodes SHOULD use passive
   acknowledgements when possible but SHOULD try requesting a network-
   layer acknowledgement one or more times before deciding that the link
   has failed and originating a Route Error to the original sender of
   the packet, as described in Section 8.3.4.

   In deciding whether or not to send a Route Error in response to
   attempting to forward a packet from some sender over a broken link, a
   node MUST limit the number of consecutive packets from a single
   sender that the node attempts to forward over this same broken link
   for which the node chooses not to return a Route Error.  This
   requirement MAY be satisfied by returning a Route Error for each
   packet that the node attempts to forward over a broken link.

8.3.1.  Using Link-Layer Acknowledgements



   If the MAC protocol in use provides feedback as to the successful
   delivery of a data packet (such as is provided for unicast packets by
   the link-layer acknowledgement frame defined by IEEE 802.11
   [IEEE80211]), then the use of the DSR Acknowledgement Request and
   Acknowledgement options is not necessary.  If such link-layer
   feedback is available, it SHOULD be used instead of any other
   acknowledgement mechanism for Route Maintenance, and the node SHOULD
   NOT
use either passive acknowledgements or network-layer
   acknowledgements for Route Maintenance.





Johnson, et al.               Experimental                     [Page 75]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   When using link-layer acknowledgements for Route Maintenance, the
   retransmission timing and the timing at which retransmission attempts
   are scheduled are generally controlled by the particular link layer
   implementation in use in the network.  For example, in IEEE 802.11,
   the link-layer acknowledgement is returned after a unicast packet as
   a part of the basic access method of the IEEE 802.11 Distributed
   Coordination Function (DCF) MAC protocol; the time at which the
   acknowledgement is expected to arrive and the time at which the next
   retransmission attempt (if necessary) will occur are controlled by
   the MAC protocol implementation.

   When a node receives a link-layer acknowledgement for any packet in
   its Maintenance Buffer, that node SHOULD remove from its Maintenance
   Buffer that packet, as well as any other packets in its Maintenance
   Buffer with the same next-hop destination.

8.3.2.  Using Passive Acknowledgements



   When link-layer acknowledgements are not available, but passive
   acknowledgements [JUBIN87] are available, passive acknowledgements
   SHOULD be used for Route Maintenance when originating or forwarding a
   packet along any hop other than the last hop (the hop leading to the
   IP Destination Address node of the packet).  In particular, passive
   acknowledgements SHOULD be used for Route Maintenance in such cases
   if the node can place its network interface into "promiscuous"
   receive mode, and if network links used for data packets generally
   operate bidirectionally.

   A node MUST NOT attempt to use passive acknowledgements for Route
   Maintenance for a packet originated or forwarded over its last hop
   (the hop leading to the IP Destination Address node of the packet),
   since the receiving node will not be forwarding the packet and thus
   no passive acknowledgement will be available to be heard by this
   node.  Beyond this restriction, a node MAY utilize a variety of
   strategies in using passive acknowledgements for Route Maintenance of
   a packet that it originates or forwards.  For example, the following
   two strategies are possible:

   -  Each time a node receives a packet to be forwarded to a node other
      than the final destination (the IP Destination Address of the
      packet), that node sends the original transmission of that packet
      without requesting a network-layer acknowledgement for it.  If no
      passive acknowledgement is received within PassiveAckTimeout after
      this transmission, the node retransmits the packet, again without
      requesting a network-layer acknowledgement for it; the same
      PassiveAckTimeout timeout value is used for each such attempt.  If
      no acknowledgement has been received after a total of
      TryPassiveAcks retransmissions of the packet, network-layer



Johnson, et al.               Experimental                     [Page 76]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      acknowledgements (as described in Section 8.3.3) are requested for
      all remaining attempts for that packet.

   -  Each node maintains a table of possible next-hop destination
      nodes, noting whether or not passive acknowledgements can
      typically be expected from transmission to that node, and the
      expected latency and jitter of a passive acknowledgement from that
      node.  Each time a node receives a packet to be forwarded to a
      node other than the IP Destination Address, the node checks its
      table of next-hop destination nodes to determine whether to use a
      passive acknowledgement or a network-layer acknowledgement for
      that transmission to that node.  The timeout for this packet can
      also be derived from this table.  A node using this method SHOULD
      prefer using passive acknowledgements to network-layer
      acknowledgements.

   In using passive acknowledgements for a packet that it originates or
   forwards, a node considers the later receipt of a new packet (e.g.,
   with promiscuous receive mode enabled on its network interface) an
   acknowledgement of this first packet if both of the following two
   tests succeed:

   -  The Source Address, Destination Address, Protocol, Identification,
      and Fragment Offset fields in the IP header of the two packets
      MUST match [RFC791].

   -  If either packet contains a DSR Source Route header, both packets
      MUST contain one, and the value in the Segments Left field in the
      DSR Source Route header of the new packet MUST be less than that
      in the first packet.

   When a node hears such a passive acknowledgement for any packet in
   its Maintenance Buffer, that node SHOULD remove from its Maintenance
   Buffer that packet, as well as any other packets in its Maintenance
   Buffer with the same next-hop destination.

8.3.3.  Using Network-Layer Acknowledgements



   When a node originates or forwards a packet and has no other
   mechanism of acknowledgement available to determine reachability of
   the next-hop node in the source route for Route Maintenance, that
   node SHOULD request a network-layer acknowledgement from that next-
   hop node.  To do so, the node inserts an Acknowledgement Request
   option in the DSR Options header in the packet.  The Identification
   field in that Acknowledgement Request option MUST be set to a value
   unique over all packets recently transmitted by this node to the same
   next-hop node.




Johnson, et al.               Experimental                     [Page 77]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   When a node receives a packet containing an Acknowledgement Request
   option, that node performs the following tests on the packet:

   -  If the indicated next-hop node address for this packet does not
      match any of this node's own IP addresses, then this node MUST NOT
      process the Acknowledgement Request option.  The indicated next-
      hop node address is the next Address[i] field in the DSR Source
      Route option in the DSR Options header in the packet, or the IP
      Destination Address in the packet if the packet does not contain a
      DSR Source Route option or the Segments Left there is zero.

   -  If the packet contains an Acknowledgement option, then this node
      MUST NOT process the Acknowledgement Request option.

   If neither of the tests above fails, then this node MUST process the
   Acknowledgement Request option by sending an Acknowledgement option
   to the previous-hop node; to do so, the node performs the following
   sequence of steps:

   -  Create a packet and set the IP Protocol field to the protocol
      number assigned for DSR (48).

   -  Set the IP Source Address field in this packet to the IP address
      of this node, copied from the source route in the DSR Source Route
      option in that packet (or from the IP Destination Address field of
      the packet, if the packet does not contain a DSR Source Route
      option).

   -  Set the IP Destination Address field in this packet to the IP
      address of the previous-hop node, copied from the source route in
      the DSR Source Route option in that packet (or from the IP Source
      Address field of the packet, if the packet does not contain a DSR
      Source Route option).

   -  Add a DSR Options header to the packet.  Set the Next Header field
      in the DSR Options header to the value 59, "No Next Header"
      [RFC2460].

   -  Add an Acknowledgement option to the DSR Options header in the
      packet; set the Acknowledgement option's Option Type field to 6
      and the Opt Data Len field to 10.

   -  Copy the Identification field from the received Acknowledgement
      Request option into the Identification field in the
      Acknowledgement option.






Johnson, et al.               Experimental                     [Page 78]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Set the ACK Source Address field in the Acknowledgement option to
      be the IP Source Address of this new packet (set above to be the
      IP address of this node).

   -  Set the ACK Destination Address field in the Acknowledgement
      option to be the IP Destination Address of this new packet (set
      above to be the IP address of the previous-hop node).

   -  Send the packet as described in Section 8.1.1.

   Packets containing an Acknowledgement option SHOULD NOT be placed in
   the Maintenance Buffer.

   When a node receives a packet with both an Acknowledgement option and
   an Acknowledgement Request option, if that node is not the
   destination of the Acknowledgement option (the IP Destination Address
   of the packet), then the Acknowledgement Request option MUST be
   ignored.  Otherwise (that node is the destination of the
   Acknowledgement option), that node MUST process the Acknowledgement
   Request option by returning an Acknowledgement option according to
   the following sequence of steps:

   -  Create a packet and set the IP Protocol field to the protocol
      number assigned for DSR (48).

   -  Set the IP Source Address field in this packet to the IP address
      of this node, copied from the source route in the DSR Source Route
      option in that packet (or from the IP Destination Address field of
      the packet, if the packet does not contain a DSR Source Route
      option).

   -  Set the IP Destination Address field in this packet to the IP
      address of the node originating the Acknowledgement option.

   -  Add a DSR Options header to the packet, and set the DSR Options
      header's Next Header field to the value 59, "No Next Header"
      [RFC2460].

   -  Add an Acknowledgement option to the DSR Options header in this
      packet; set the Acknowledgement option's Option Type field to 6
      and the Opt Data Len field to 10.

   -  Copy the Identification field from the received Acknowledgement
      Request option into the Identification field in the
      Acknowledgement option.






Johnson, et al.               Experimental                     [Page 79]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Set the ACK Source Address field in the option to the IP Source
      Address of this new packet (set above to be the IP address of this
      node).

   -  Set the ACK Destination Address field in the option to the IP
      Destination Address of this new packet (set above to be the IP
      address of the node originating the Acknowledgement option).

   -  Send the packet directly to the destination.  The IP Destination
      Address MUST be treated as a direct neighbor node: the
      transmission to that node MUST be done in a single IP forwarding
      hop, without Route Discovery and without searching the Route
      Cache.  In addition, this packet MUST NOT contain a DSR
      Acknowledgement Request, MUST NOT be retransmitted for Route
      Maintenance, and MUST NOT expect a link-layer acknowledgement or
      passive acknowledgement.

   When using network-layer acknowledgements for Route Maintenance, a
   node SHOULD use an adaptive algorithm in determining the
   retransmission timeout for each transmission attempt of an
   acknowledgement request.  For example, a node SHOULD maintain a
   separate round-trip time (RTT) estimate for each node to which it has
   recently attempted to transmit packets, and it SHOULD use this RTT
   estimate in setting the timeout for each retransmission attempt for
   Route Maintenance.  The TCP RTT estimation algorithm has been shown
   to work well for this purpose in implementation and testbed
   experiments with DSR [MALTZ99b, MALTZ01].

8.3.4.  Originating a Route Error



   When a node is unable to verify reachability of a next-hop node after
   reaching a maximum number of retransmission attempts, it SHOULD send
   a Route Error to the IP Source Address of the packet.  When sending a
   Route Error for a packet containing either a Route Error option or an
   Acknowledgement option, a node SHOULD add these existing options to
   its Route Error, subject to the limit described below.

   A node transmitting a Route Error MUST perform the following steps:

   -  Create an IP packet and set the IP Protocol field to the protocol
      number assigned for DSR (48).  Set the Source Address field in
      this packet's IP header to the address of this node.

   -  If the Salvage field in the DSR Source Route option in the packet
      triggering the Route Error is zero, then copy the Source Address
      field of the packet triggering the Route Error into the
      Destination Address field in the new packet's IP header;




Johnson, et al.               Experimental                     [Page 80]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      otherwise, copy the Address[1] field from the DSR Source Route
      option of the packet triggering the Route Error into the
      Destination Address field in the new packet's IP header

   -  Insert a DSR Options header into the new packet.

   -  Add a Route Error Option to the new packet, setting the Error Type
      to NODE_UNREACHABLE, the Salvage value to the Salvage value from
      the DSR Source Route option of the packet triggering the Route
      Error, and the Unreachable Node Address field to the address of
      the next-hop node from the original source route.  Set the Error
      Source Address field to this node's IP address, and the Error
      Destination field to the new packet's IP Destination Address.

   -  If the packet triggering the Route Error contains any Route Error
      or Acknowledgement options, the node MAY append to its Route Error
      each of these options, with the following constraints:

      o  The node MUST NOT include any Route Error option from the
         packet triggering the new Route Error, for which the total
         Salvage count (Section 6.4) of that included Route Error would
         be greater than MAX_SALVAGE_COUNT in the new packet.

      o  If any Route Error option from the packet triggering the new
         Route Error is not included in the packet, the node MUST NOT
         include any following Route Error or Acknowledgement options
         from the packet triggering the new Route Error.

      o  Any appended options from the packet triggering the Route Error
         MUST follow the new Route Error in the packet.

      o  In appending these options to the new Route Error, the order of
         these options from the packet triggering the Route Error MUST
         be preserved.

   -  Send the packet as described in Section 8.1.1.

8.3.5.  Processing a Received Route Error Option



   When a node receives a packet containing a Route Error option, that
   node MUST process the Route Error option according to the following
   sequence of steps:

   -  The node MUST remove from its Route Cache the link from the node
      identified by the Error Source Address field to the node
      identified by the Unreachable Node Address field (if this link is
      present in its Route Cache).  If the node implements its Route
      Cache as a link cache, as described in Section 4.1, only this



Johnson, et al.               Experimental                     [Page 81]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      single link is removed; if the node implements its Route Cache as
      a path cache, however, all routes (paths) that use this link are
      either truncated before the link or removed completely.

   -  If the option following the Route Error is an Acknowledgement or
      Route Error option sent by this node (that is, with
      Acknowledgement or Error Source Address equal to this node's
      address), copy the DSR options following the current Route Error
      into a new packet with IP Source Address equal to this node's own
      IP address and IP Destination Address equal to the Acknowledgement
      or Error Destination Address.  Transmit this packet as described
      in Section 8.1.1, with the Salvage count in the DSR Source Route
      option set to the Salvage value of the Route Error.

   In addition, after processing the Route Error as described above, the
   node MAY initiate a new Route Discovery for any destination node for
   which it then has no route in its Route Cache as a result of
   processing this Route Error, if the node has indication that a route
   to that destination is needed.  For example, if the node has an open
   TCP connection to some destination node, then if the processing of
   this Route Error removed the only route to that destination from this
   node's Route Cache, then this node MAY initiate a new Route Discovery
   for that destination node.  Any node, however, MUST limit the rate at
   which it initiates new Route Discoveries for any single destination
   address, and any new Route Discovery initiated in this way as part of
   processing this Route Error MUST conform as a part of this limit.

8.3.6.  Salvaging a Packet



   When an intermediate node forwarding a packet detects through Route
   Maintenance that the next-hop link along the route for that packet is
   broken (Section 8.3), if the node has another route to the packet's
   IP Destination Address in its Route Cache, the node SHOULD "salvage"
   the packet rather than discard it.  To do so using the route found in
   its Route Cache, this node processes the packet as follows:

   -  If the MAC protocol in use in the network is not capable of
      transmitting unicast packets over unidirectional links, as
      discussed in Section 3.3.1, then if this packet contains a Route
      Reply option, remove and discard the Route Reply option in the
      packet; if the DSR Options header in the packet then contains no
      DSR options or only a DSR Source Route Option, remove the DSR
      Options header from the packet.  If the resulting packet then
      contains only an IP header (e.g., no transport layer header or
      payload), the node SHOULD NOT salvage the packet and instead
      SHOULD discard the entire packet.





Johnson, et al.               Experimental                     [Page 82]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Modify the existing DSR Source Route option in the packet so that
      the Address[i] fields represent the source route found in this
      node's Route Cache to this packet's IP Destination Address.
      Specifically, the node copies the hop addresses of the source
      route into sequential Address[i] fields in the DSR Source Route
      option, for i = 1, 2, ..., n.  Address[1], here, is the address of
      the salvaging node itself (the first address in the source route
      found from this node to the IP Destination Address of the packet).
      The value n, here, is the number of hop addresses in this source
      route, excluding the destination of the packet (which is instead
      already represented in the Destination Address field in the
      packet's IP header).

   -  Initialize the Segments Left field in the DSR Source Route option
      to n as defined above.

   -  The First Hop External (F) bit in the DSR Source Route option MUST
      be set to 0.

   -  The Last Hop External (L) bit in the DSR Source Route option is
      copied from the External bit flagging the last hop in the source
      route for the packet, as indicated in the Route Cache.

   -  The Salvage field in the DSR Source Route option is set to 1 plus
      the value of the Salvage field in the DSR Source Route option of
      the packet that caused the error.

   -  Transmit the packet to the next-hop node on the new source route
      in the packet, using the forwarding procedure described in Section
      8.1.5.

   As described in Section 8.3.4, the node in this case also SHOULD
   return a Route Error to the original sender of the packet.  If the
   node chooses to salvage the packet, it SHOULD do so after originating
   the Route Error.

   When returning any Route Reply in the case in which the MAC protocol
   in use in the network is not capable of transmitting unicast packets
   over unidirectional links, the source route used for routing the
   Route Reply packet MUST be obtained by reversing the sequence of hops
   in the Route Request packet (the source route that is then returned
   in the Route Reply).  This restriction on returning a Route Reply and
   on salvaging a packet that contains a Route Reply option enables the
   Route Reply to test this sequence of hops for bidirectionality,
   preventing the Route Reply from being received by the initiator of
   the Route Discovery unless each of the hops over which the Route
   Reply is returned (and thus each of the hops in the source route
   being returned in the Reply) is bidirectional.



Johnson, et al.               Experimental                     [Page 83]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


8.4.  Multiple Network Interface Support



   A node using DSR MAY have multiple network interfaces that support
   DSR ad hoc network routing.  This section describes special packet
   processing at such nodes.

   A node with multiple network interfaces that support DSR ad hoc
   network routing MUST have some policy for determining which Route
   Request packets are forwarded using which network interfaces.  For
   example, a node MAY choose to forward all Route Requests over all
   network interfaces.

   When a node with multiple network interfaces that support DSR
   propagates a Route Request on a network interface other than the one
   on which it received the Route Request, it MUST in this special case
   modify the Address list in the Route Request as follows:

   -  Append the node's IP address for the incoming network interface.

   -  Append the node's IP address for the outgoing network interface.

   When a node forwards a packet containing a source route, it MUST
   assume that the next-hop node is reachable on the incoming network
   interface, unless the next hop is the address of one of this node's
   network interfaces, in which case this node MUST skip over this
   address in the source route and process the packet in the same way as
   if it had just received it from that network interface, as described
   in Section 8.1.5.

   If a node that previously had multiple network interfaces that
   support DSR receives a packet sent with a source route specifying a
   change to a network interface, as described above, that is no longer
   available, it MAY send a Route Error to the source of the packet
   without attempting to forward the packet on the incoming network
   interface, unless the network uses an autoconfiguration mechanism
   that may have allowed another node to acquire the now unused address
   of the unavailable network interface.

8.5.  IP Fragmentation and Reassembly



   When a node using DSR wishes to fragment a packet that contains a DSR
   header not containing a Route Request option, it MUST perform the
   following sequence of steps:

   -  Remove the DSR Options header from the packet.






Johnson, et al.               Experimental                     [Page 84]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   -  Fragment the packet using normal IP fragmentation processing
      [RFC791].  However, when determining the size of each fragment to
      create from the original packet, the fragment size MUST be reduced
      by the size of the DSR Options header from the original packet.

   -  IP-in-IP encapsulate each fragment [RFC2003].  The IP Destination
      address of the outer (encapsulating) packet MUST be set equal to
      the IP Destination address of the original packet.

   -  Add the DSR Options header from the original packet to each
      resulting encapsulating packet.  If a Source Route header is
      present in the DSR Options header, increment the Salvage field.

   When a node using the DSR protocol receives an IP-in-IP encapsulated
   packet destined to itself, it SHOULD decapsulate the packet [RFC2003]
   and then process the inner packet according to standard IP reassembly
   processing [RFC791].

8.6.  Flow State Processing



   A node implementing the optional DSR flow state extension MUST follow
   these additional processing steps.

8.6.1.  Originating a Packet



   When originating any packet to be routed using flow state, a node
   using DSR flow state MUST do the following:

   -  If the route to be used for this packet has never had a DSR flow
      state established along it (or the existing flow state has
      expired):

      o  Generate a 16-bit Flow ID larger than any unexpired Flow IDs
         used by this node for this destination.  Odd Flow IDs MUST be
         chosen for "default" flows; even Flow IDs MUST be chosen for
         non-default flows.

      o  Add a DSR Options header, as described in Section 8.1.2.

      o  Add a DSR Flow State header, as described in Section 8.6.2.

      o  Initialize the Hop Count field in the DSR Flow State header to
         0.

      o  Set the Flow ID field in the DSR Flow State header to the Flow
         ID generated in the first step.

      o  Add a Timeout option to the DSR Options header.



Johnson, et al.               Experimental                     [Page 85]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  Add a Source Route option after the Timeout option with the
         route to be used, as described in Section 8.1.3.

      o  The source node SHOULD record this flow in its Flow Table.

      o  If this flow is recorded in the Flow Table, the TTL in this
         Flow Table entry MUST be set to be the TTL of this flow
         establishment packet.

      o  If this flow is recorded in the Flow Table, the timeout in this
         Flow Table entry MUST be set to a value no less than the value
         specified in the Timeout option.

   -  If the route to be used for this packet has had DSR flow state
      established along it, but has not been established end-to-end:

      o  Add a DSR Options header, as described in Section 8.1.2.

      o  Add a DSR Flow State header, as described in Section 8.6.2.

      o  Initialize the Hop Count field in the DSR Flow State header to
         0.

      o  The Flow ID field of the DSR Flow State header SHOULD be the
         Flow ID previously used for this route.  If it is not, the
         steps for sending packets along never-before-established routes
         above MUST be followed in place of these.

      o  Add a Timeout option to the DSR Options header, setting the
         Timeout to a value not greater than the timeout remaining for
         this flow in the Flow Table.

      o  Add a Source Route option after the Timeout option with the
         route to be used, as described in Section 8.1.3.

      o  If the IP TTL is not equal to the TTL specified in the Flow
         Table, the source node MUST set a flag to indicate that this
         flow cannot be used as default.

   -  If the route the node wishes to use for this packet has been
      established as a flow end-to-end and is not the default flow:

      o  Add a DSR Flow State header, as described in Section 8.6.2.

      o  Initialize the Hop Count field in the DSR Flow State header to
         0.





Johnson, et al.               Experimental                     [Page 86]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  The Flow ID field of the DSR Flow State header SHOULD be set to
         the Flow ID previously used for this route.  If it is not, the
         steps for sending packets along never-before-established routes
         above MUST be followed in place of these.

      o  If the next hop requires a network-layer acknowledgement for
         Route Maintenance, add a DSR Options header, as described in
         Section 8.1.2, and an Acknowledgement Request option, as
         described in Section 8.3.3.

      o  A DSR Options header SHOULD NOT be added to a packet, unless it
         is added to carry an Acknowledgement Request option, in which
         case:

         +  A Source Route option in the DSR Options header SHOULD NOT
            be added.

         +  If a Source Route option in the DSR Options header is added,
            the steps for sending packets along flows not yet
            established end-to-end MUST be followed in place of these.

         +  A Timeout option SHOULD NOT be added.

         +  If a Timeout option is added, it MUST specify a timeout not
            greater than the timeout remaining for this flow in the Flow
            Table.

   -  If the route the node wishes to use for this packet has been
      established as a flow end-to-end and is the current default flow:

      o  If the IP TTL is not equal to the TTL specified in the Flow
         Table, the source node MUST follow the steps above for sending
         a packet along a non-default flow that has been established
         end-to-end in place of these steps.

      o  If the next hop requires a network-layer acknowledgement for
         Route Maintenance, the sending node MUST add a DSR Options
         header and an Acknowledgement Request option, as described in
         Section 8.3.3.  The sending node MUST NOT add any additional
         options to this header.

      o  A DSR Options header SHOULD NOT be added, except as specified
         in the previous step.  If one is added in a way inconsistent
         with the previous step, the source node MUST follow the steps
         above for sending a packet along a non-default flow that has
         been established end-to-end in place of these steps.





Johnson, et al.               Experimental                     [Page 87]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


8.6.2.  Inserting a DSR Flow State Header



   A node originating a packet adds a DSR Flow State header to the
   packet, if necessary, to carry information needed by the routing
   protocol.  A packet MUST NOT contain more than one DSR Flow State
   header.  A DSR Flow State header is added to a packet by performing
   the following sequence of steps:

   -  Insert a DSR Flow State header after the IP header and any Hop-
      by-Hop Options header that may already be in the packet, but
      before any other header that may be present.

   -  Set the Next Header field of the DSR Flow State header to the Next
      Header field of the previous header (either an IP header or a
      Hop-by-Hop Options header).

   -  Set the Flow (F) bit in the DSR Flow State header to 1.

   -  Set the Protocol field of the IP header to the protocol number
      assigned for DSR (48).

8.6.3.  Receiving a Packet



   This section describes processing only for packets that are sent to
   this processing node as the next-hop node; that is, when the MAC-
   layer destination address is the MAC address of this node.
   Otherwise, the process described in Sections 8.6.5 should be
   followed.

   The flow along which a packet is being sent is considered to be in
   the Flow Table if the triple (IP Source Address, IP Destination
   Address, Flow ID) has an unexpired entry in this node's Flow Table.

   When a node using DSR flow state receives a packet, it MUST follow
   the following steps for processing:

   -  If a DSR Flow State header is present, increment the Hop Count
      field.

   -  In addition, if a DSR Flow State header is present, then if the
      triple (IP Source Address, IP Destination Address, Flow ID) is in
      this node's Automatic Route Shortening Table and the packet is
      listed in the entry, then the node MAY send a gratuitous Route
      Reply as described in Section 4.4, subject to the rate limiting
      specified therein.  This gratuitous Route Reply gives the route by
      which the packet originally reached this node.  Specifically, the
      node sending the gratuitous Route Reply constructs the route to
      return in the Route Reply as follows:



Johnson, et al.               Experimental                     [Page 88]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  Let k = (packet Hop Count) - (table Hop Count), where packet
         Hop Count is the value of the Hop Count field in this received
         packet, and table Hop Count is the Hop Count value stored for
         this packet in the corresponding entry in this node's Automatic
         Route Shortening Table.

      o  Copy the complete source route for this flow from the
         corresponding entry in the node's Flow Table.

      o  Remove from this route the k hops immediately preceding this
         node in the route, since these are the hops "skipped over" by
         the packet as recorded in the Automatic Route Shortening Table
         entry.

   -  Process each of the DSR options within the DSR Options header in
      order:

      o  On receiving a Pad1 or PadN option, skip over the option.

      o  On receiving a Route Request for which this node is the
         destination, remove the option and return a Route Reply as
         specified in Section 8.2.2.

      o  On receiving a broadcast Route Request that this node has not
         previously seen for which this node is not the destination,
         append this node's incoming interface address to the Route
         Request, continue propagating the Route Request as specified in
         Section 8.2.2, pass the payload, if any, to the network layer,
         and stop processing.

      o  On receiving a Route Request that this node has previously seen
         for which this node is not the destination, discard the packet
         and stop processing.

      o  On receiving any Route Request, add appropriate links to the
         Route Cache, as specified in Section 8.2.2.

      o  On receiving a Route Reply for which this node is the
         initiator, remove the Route Reply from the packet and process
         it as specified in Section 8.2.6.

      o  On receiving any Route Reply, add appropriate links to the
         Route Cache, as specified in Section 8.2.6.

      o  On receiving any Route Error of type NODE_UNREACHABLE, remove
         appropriate links to the Route Cache, as specified in Section
         8.3.5.




Johnson, et al.               Experimental                     [Page 89]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  On receiving a Route Error of type NODE_UNREACHABLE that this
         node is the Error Destination Address of, remove the Route
         Error from the packet and process it as specified in Section
         8.3.5.  It also MUST stop originating packets along any flows
         using the link from Error Source Address to Unreachable Node,
         and it MAY remove from its Flow Table any flows using the link
         from Error Source Address to Unreachable Node.

      o  On receiving a Route Error of type UNKNOWN_FLOW that this node
         is not the Error Destination Address of, the node checks if the
         Route Error corresponds to a flow in its Flow Table.  If it
         does not, the node silently discards the Route Error;
         otherwise, it forwards the packet to the expected previous hop
         of the corresponding flow.  If Route Maintenance cannot confirm
         the reachability of the previous hop, the node checks if the
         network interface requires bidirectional links for operation.
         If it does, the node silently discards the Route Error;
         otherwise, it sends the Error as if it were originating it, as
         described in Section 8.1.1.

      o  On receiving a Route Error of type UNKNOWN_FLOW that this node
         is the Error Destination Address of, remove the Route Error
         from the packet and mark the flow specified by the triple
         (Error Destination Address, Original IP Destination Address,
         Flow ID) as not having been established end-to-end.

      o  On receiving a Route Error of type DEFAULT_FLOW_UNKNOWN that
         this node is not the Error Destination Address of, the node
         checks if the Route Error corresponds to a flow in its Default
         Flow Table.  If it does not, the node silently discards the
         Route Error; otherwise, it forwards the packet to the expected
         previous hop of the corresponding flow.  If Route Maintenance
         cannot confirm the reachability of the previous hop, the node
         checks if the network interface requires bidirectional links
         for operation.  If it does, the node silently discards the
         Route Error; otherwise, it sends the Error as if it were
         originating it, as described in Section 8.1.1.

      o  On receiving a Route Error of type DEFAULT_FLOW_UNKNOWN that
         this node is the Error Destination Address of, remove the Route
         Error from the packet and mark the default flow between the
         Error Destination Address and the Original IP Destination
         Address as not having been established end-to-end.








Johnson, et al.               Experimental                     [Page 90]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  On receiving an Acknowledgement Request option, the receiving
         node removes the Acknowledgement Request option and replies to
         the previous hop with an Acknowledgement option.  If the
         previous hop cannot be determined, the Acknowledgement Request
         option is discarded, and processing continues.

      o  On receiving an Acknowledgement option, the receiving node
         removes the Acknowledgement option and processes it.

      o  On receiving any Acknowledgement option, add the appropriate
         link to the Route Cache, as specified in Section 8.1.4.

      o  On receiving any Source Route option, add appropriate links to
         the Route Cache, as specified in Section 8.1.4.

      o  On receiving a Source Route option, if no DSR Flow State header
         is present, if the flow this packet is being sent along is in
         the Flow Table, or if no Timeout option preceded the Source
         Route option in this DSR Options header, process it as
         specified in Section 8.1.4.  Stop processing this packet unless
         the last address in the Source Route option is an address of
         this node.

      o  On receiving a Source Route option in a packet with a DSR Flow
         State header, if the Flow ID specified in the DSR Flow State
         header is not in the Flow Table, add the flow to the Flow
         Table, setting the Timeout value to a value not greater than
         the Timeout field of the Timeout option in this header.  If no
         Timeout option preceded the Source Route option in this header,
         the flow MUST NOT be added to the Flow Table.

         If the Flow ID is odd and larger than any unexpired, odd Flow
         IDs for this (IP Source Address, IP Destination Address), it is
         set to be default in the Default Flow ID Table.

         Then process the Route option as specified in Section 8.1.4.
         Stop processing this packet unless the last address in the
         Source Route option is an address of this node.

      o  On receiving a Timeout option, check if this packet contains a
         DSR Flow State header.  If this packet does not contain a DSR
         Flow State header, discard the DSR option.  Otherwise, record
         the Timeout value in the option for future reference.  The
         value recorded SHOULD be discarded when the node has finished
         processing this DSR Options header.  If the flow that this
         packet is being sent along is in the Flow Table, it MAY set the
         flow to time out no more than Timeout seconds in the future.




Johnson, et al.               Experimental                     [Page 91]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


      o  On receiving a Destination and Flow ID option, if the IP
         Destination Address is not an address of this node, forward the
         packet according to the Flow ID, as described in Section 8.6.4,
         and stop processing this packet.

      o  On receiving a Destination and Flow ID option, if the IP
         Destination Address is an address of this node, set the IP
         Destination Address to the New IP Destination Address specified
         in the option and set the Flow ID to the New Flow Identifier.
         Then remove the Destination and Flow ID option from the packet
         and continue processing.

   -  If the IP Destination Address is an address of this node, remove
      the DSR Options header, if any, pass the packet up the network
      stack, and stop processing.

   -  If there is still a DSR Options header containing no options,
      remove the DSR Options header.

   -  If there is still a DSR Flow State header, forward the packet
      according to the Flow ID, as described in Section 8.6.4.

   -  If there is neither a DSR Options header nor a DSR Flow State
      header, but there is an entry in the Default Flow Table for the
      (IP Source Address, IP Destination Address) pair:

      o  If the IP TTL is not equal to the TTL expected in the Flow
         Table, insert a DSR Flow State header, setting the Hop Count
         equal to the Hop Count of this node, and the Flow ID equal to
         the default Flow ID found in the Default Flow Table, and
         forward this packet according to the Flow ID, as described in
         Section 8.6.4.

      o  Otherwise, follow the steps for forwarding the packet using
         Flow IDs described in Section 8.6.4, but taking the Flow ID to
         be the default Flow ID found in the Default Flow Table.

   -  If there is no DSR Options header and no DSR Flow State header and
      no default flow can be found, the node returns a Route Error of
      type DEFAULT_FLOW_UNKNOWN to the IP Source Address, specifying the
      IP Destination Address as the Original IP Destination in the
      type-specific field.









Johnson, et al.               Experimental                     [Page 92]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


8.6.4.  Forwarding a Packet Using Flow IDs



   To forward a packet using Flow IDs, a node MUST follow the following
   sequence of steps:

   -  If the triple (IP Source Address, IP Destination Address, Flow ID)
      is not in the Flow Table, return a Route Error of type
      UNKNOWN_FLOW.

   -  If a network-layer acknowledgement is required for Route
      Maintenance for the next hop, the node MUST include an
      Acknowledgement Request option as specified in Section 8.3.3.  If
      no DSR Options header is in the packet in which the
      Acknowledgement Request option is to be added, it MUST be
      included, as described in Section 8.1.2, except that it MUST be
      added after the DSR Flow State header, if one is present.

   -  Attempt to transmit this packet to the next hop as specified in
      the Flow Table, performing Route Maintenance to detect broken
      routes.

8.6.5.  Promiscuously Receiving a Packet



   This section describes processing only for packets that have MAC
   destinations other than this processing node.  Otherwise, the process
   described in Section 8.6.3 should be followed.

   When a node using DSR flow state promiscuously overhears a packet, it
   SHOULD follow the following steps for processing:

   -  If the packet contains a DSR Flow State header, and if the triple
      (IP Source Address, IP Destination Address, Flow ID) is in the
      Flow Table and the Hop Count is less than the Hop Count in the
      flow's entry, the node MAY retain the packet in the Automatic
      Route Shortening Table.  If it can be determined that this Flow ID
      has been recently used, the node SHOULD retain the packet in the
      Automatic Route Shortening Table.

   -  If the packet contains neither a DSR Flow State header nor a
      Source Route option and a Default Flow ID can be found in the
      Default Flow Table for the (IP Source Address, IP Destination
      Address), and if the IP TTL is greater than the TTL in the Flow
      Table for the default flow, the node MAY retain the packet in the
      Automatic Route Shortening Table.  If it can be determined that
      this Flow ID has been used recently, the node SHOULD retain the
      packet in the Automatic Route Shortening Table.





Johnson, et al.               Experimental                     [Page 93]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


8.6.6.  Operation Where the Layer below DSR Decreases the IP TTL
        Non-uniformly



   Some nodes may use an IP tunnel as a DSR hop.  If different packets
   sent along this IP tunnel can take different routes, the reduction in
   IP TTL across this link may be different for different packets.  This
   prevents the Automatic Route Shortening and Loop Detection
   functionality from working properly when used in conjunction with
   default routes.

   Nodes forwarding packets without a Source Route option onto a link
   with unpredictable TTL changes MUST ensure that a DSR Flow State
   header is present, indicating the correct Hop Count and Flow ID.

8.6.7.  Salvage Interactions with DSR



   Nodes salvaging packets MUST remove the DSR Flow State header, if
   present.

   Anytime this document refers to the Salvage field in the Source Route
   option, packets without a Source Route option are considered to have
   the value zero in the Salvage field.





























Johnson, et al.               Experimental                     [Page 94]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


9.  Protocol Constants and Configuration Variables



   Any DSR implementation MUST support the following configuration
   variables and MUST support a mechanism enabling the value of these
   variables to be modified by system management.  The specific variable
   names are used for demonstration purposes only, and an implementation
   is not required to use these names for the configuration variables,
   so long as the external behavior of the implementation is consistent
   with that described in this document.

   For each configuration variable below, the default value is specified
   to simplify configuration.  In particular, the default values given
   below are chosen for a DSR network running over 2 Mbps IEEE 802.11
   network interfaces using the Distributed Coordination Function (DCF)
   MAC protocol with RTS and CTS [IEEE80211, BROCH98].

      DiscoveryHopLimit                  255   hops

      BroadcastJitter                     10   milliseconds

      RouteCacheTimeout                  300   seconds

      SendBufferTimeout                   30   seconds

      RequestTableSize                    64   nodes
      RequestTableIds                     16   identifiers
      MaxRequestRexmt                     16   retransmissions
      MaxRequestPeriod                    10   seconds
      RequestPeriod                      500   milliseconds
      NonpropRequestTimeout               30   milliseconds

      RexmtBufferSize                     50   packets

      MaintHoldoffTime                   250   milliseconds

      MaxMaintRexmt                        2   retransmissions

      TryPassiveAcks                       1   attempt
      PassiveAckTimeout                  100   milliseconds

      GratReplyHoldoff                     1   second

   In addition, the following protocol constant MUST be supported by any
   implementation of the DSR protocol:

      MAX_SALVAGE_COUNT                   15   salvages





Johnson, et al.               Experimental                     [Page 95]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


10.  IANA Considerations



   This document specifies the DSR Options header and DSR Flow State
   header, for which the IP protocol number 48 has been assigned.  A
   single IP protocol number can be used for both header types, since
   they can be distinguished by the Flow State Header (F) bit in each
   header.

   In addition, this document proposes use of the value "No Next Header"
   (originally defined for use in IPv6 [RFC2460]) within an IPv4 packet,
   to indicate that no further header follows a DSR Options header.

   Finally, this document introduces a number of DSR options for use in
   the DSR Options header, and additional new DSR options may be defined
   in the future.  Each of these options requires a unique Option Type
   value, the most significant 3 bits (that is, Option Type & 0xE0)
   encoded as defined in Section 6.1.  It is necessary only that each
   Option Type value be unique, not that they be unique in the remaining
   5 bits of the value after these 3 most significant bits.

   Two registries (DSR Protocol Options and DSR Protocol Route Error
   Types) have been created and contain the initial registrations.
   Assignment of new values for DSR options will be by Expert Review
   [RFC2434], with the authors of this document serving as the
   Designated Experts.

11.  Security Considerations



   This document does not specifically address security concerns.  This
   document does assume that all nodes participating in the DSR protocol
   do so in good faith and without malicious intent to corrupt the
   routing ability of the network.

   Depending on the threat model, a number of different mechanisms can
   be used to secure DSR.  For example, in an environment where node
   compromise is unrealistic and where all the nodes participating in
   the DSR protocol share a common goal that motivates their
   participation in the protocol, the communications between the nodes
   can be encrypted at the physical channel or link layer to prevent
   attack by outsiders.  Cryptographic approaches, such as that provided
   by Ariadne [HU02] or Secure Routing Protocol (SRP)
   [PAPADIMITRATOS02], can resist stronger attacks.









Johnson, et al.               Experimental                     [Page 96]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Appendix A.  Link-MaxLife Cache Description



   As guidance to implementers of DSR, the description below outlines
   the operation of a possible implementation of a Route Cache for DSR
   that has been shown to outperform other caches studied in detailed
   simulations.  Use of this design for the Route Cache is recommended
   in implementations of DSR.

   This cache, called "Link-MaxLife" [HU00], is a link cache, in that
   each individual link (hop) in the routes returned in Route Reply
   packets (or otherwise learned from the header of overhead packets) is
   added to a unified graph data structure of this node's current view
   of the network topology, as described in Section 4.1.  To search for
   a route in this cache to some destination node, the sending node uses
   a graph search algorithm, such as the well-known Dijkstra's
   shortest-path algorithm, to find the current best path through the
   graph to the destination node.

   The Link-MaxLife form of link cache is adaptive in that each link in
   the cache has a timeout that is determined dynamically by the caching
   node according to its observed past behavior of the two nodes at the
   ends of the link; in addition, when selecting a route for a packet
   being sent to some destination, among cached routes of equal length
   (number of hops) to that destination, Link-MaxLife selects the route
   with the longest expected lifetime (highest minimum timeout of any
   link in the route).

   Specifically, in Link-MaxLife, a link's timeout in the Route Cache is
   chosen according to a "Stability Table" maintained by the caching
   node.  Each entry in a node's Stability Table records the address of
   another node and a factor representing the perceived "stability" of
   this node.  The stability of each other node in a node's Stability
   Table is initialized to InitStability.  When a link from the Route
   Cache is used in routing a packet originated or salvaged by that
   node, the stability metric for each of the two endpoint nodes of that
   link is incremented by the amount of time since that link was last
   used, multiplied by StabilityIncrFactor (StabilityIncrFactor >= 1);
   when a link is observed to break and the link is thus removed from
   the Route Cache, the stability metric for each of the two endpoint
   nodes of that link is multiplied by StabilityDecrFactor
   (StabilityDecrFactor < 1).

   When a node adds a new link to its Route Cache, the node assigns a
   lifetime for that link in the Cache equal to the stability of the
   less "stable" of the two endpoint nodes for the link, except that a
   link is not allowed to be given a lifetime less than MinLifetime.
   When a link is used in a route chosen for a packet originated or
   salvaged by this node, the link's lifetime is set to be at least



Johnson, et al.               Experimental                     [Page 97]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   UseExtends into the future; if the lifetime of that link in the Route
   Cache is already further into the future, the lifetime remains
   unchanged.

   When a node using Link-MaxLife selects a route from its Route Cache
   for a packet being originated or salvaged by this node, it selects
   the shortest-length route that has the longest expected lifetime
   (highest minimum timeout of any link in the route), as opposed to
   simply selecting an arbitrary route of shortest length.

   The following configuration variables are used in the description of
   Link-MaxLife above.  The specific variable names are used for
   demonstration purposes only, and an implementation is not required to
   use these names for these configuration variables.  For each
   configuration variable below, the default value is specified to
   simplify configuration.  In particular, the default values given
   below are chosen for a DSR network where nodes move at relative
   velocities between 12 and 25 seconds per wireless transmission
   radius.

      InitStability                       25   seconds
      StabilityIncrFactor                  4
      StabilityDecrFactor                0.5

      MinLifetime                          1   second
      UseExtends                         120   seconds

























Johnson, et al.               Experimental                     [Page 98]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Appendix B.  Location of DSR in the ISO Network Reference Model



   When designing DSR, we had to determine at what layer within the
   protocol hierarchy to implement ad hoc network routing.  We
   considered two different options: routing at the link layer (ISO
   layer 2) and routing at the network layer (ISO layer 3).  Originally,
   we opted to route at the link layer for several reasons:

   -  Pragmatically, running the DSR protocol at the link layer
      maximizes the number of mobile nodes that can participate in ad
      hoc networks.  For example, the protocol can route equally well
      between IPv4 [RFC791], IPv6 [RFC2460], and IPX [TURNER90] nodes.

   -  Historically [JOHNSON94, JOHNSON96a], DSR grew from our
      contemplation of a multi-hop propagating version of the Internet's
      Address Resolution Protocol (ARP) [RFC826], as well as from the
      routing mechanism used in IEEE 802 source routing bridges
      [PERLMAN92].  These are layer 2 protocols.

   -  Technically, we designed DSR to be simple enough that it could be
      implemented directly in the firmware inside wireless network
      interface cards [JOHNSON94, JOHNSON96a], well below the layer 3
      software within a mobile node.  We see great potential in this for
      DSR running inside a cloud of mobile nodes around a fixed base
      station, where DSR would act to transparently extend the coverage
      range to these nodes.  Mobile nodes that would otherwise be unable
      to communicate with the base station due to factors such as
      distance, fading, or local interference sources could then reach
      the base station through their peers.

   Ultimately, however, we decided to specify and to implement
   [MALTZ99b] DSR as a layer 3 protocol, since this is the only layer at
   which we could realistically support nodes with multiple network
   interfaces of different types forming an ad hoc network.

















Johnson, et al.               Experimental                     [Page 99]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Appendix C.  Implementation and Evaluation Status



   The initial design of the DSR protocol, including DSR's basic Route
   Discovery and Route Maintenance mechanisms, was first published in
   December 1994 [JOHNSON94]; significant additional design details and
   initial simulation results were published in early 1996 [JOHNSON96a].

   The DSR protocol has been extensively studied since then through
   additional detailed simulations.  In particular, we have implemented
   DSR in the ns-2 network simulator [NS-2, BROCH98] and performed
   extensive simulations of DSR using ns-2 (e.g., [BROCH98, MALTZ99a]).
   We have also conducted evaluations of the different caching
   strategies in this document [HU00].

   We have also implemented the DSR protocol under the FreeBSD 2.2.7
   operating system running on Intel x86 platforms.  FreeBSD [FREEBSD]
   is based on a variety of free software, including 4.4 BSD Lite, from
   the University of California, Berkeley.  For the environments in
   which we used it, this implementation is functionally equivalent to
   the version of the DSR protocol specified in this document.

   During the 7 months from August 1998 to February 1999, we designed
   and implemented a full-scale physical testbed to enable the
   evaluation of ad hoc network performance in the field, in an actively
   mobile ad hoc network under realistic communication workloads.  The
   last week of February and the first week of March of 1999 included
   demonstrations of this testbed to a number of our sponsors and
   partners, including Lucent Technologies, Bell Atlantic, and the
   Defense Advanced Research Projects Agency (DARPA).  A complete
   description of the testbed is available [MALTZ99b, MALTZ00, MALTZ01].

   We have since ported this implementation of DSR to FreeBSD 3.3, and
   we have also added a preliminary version of Quality of Service (QoS)
   support for DSR.  A demonstration of this modified version of DSR was
   presented in July 2000.  These QoS features are not included in this
   document and will be added later in a separate document on top of the
   base protocol specified here.

   DSR has also been implemented under Linux by Alex Song at the
   University of Queensland, Australia [SONG01].  This implementation
   supports the Intel x86 PC platform and the Compaq iPAQ.

   The Network and Telecommunications Research Group at Trinity College,
   Dublin, have implemented a version of DSR on Windows CE.

   Microsoft Research has implemented a version of DSR on Windows XP and
   has used it in testbeds of over 15 nodes.  Several machines use this
   implementation as their primary means of accessing the Internet.



Johnson, et al.               Experimental                    [Page 100]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   Several other independent groups have also used DSR as a platform for
   their own research, or as a basis of comparison between ad hoc
   network routing protocols.

   A preliminary version of the optional DSR flow state extension was
   implemented in FreeBSD 3.3.  A demonstration of this modified version
   of DSR was presented in July 2000.  The DSR flow state extension has
   also been extensively evaluated using simulation [HU01].

Acknowledgements



   The protocol described in this document has been designed and
   developed within the Monarch Project, a long-term research project at
   Rice University (previously at Carnegie Mellon University) that is
   developing adaptive networking protocols and protocol interfaces to
   allow truly seamless wireless and mobile node networking [JOHNSON96b,
   MONARCH].

   The authors would like to acknowledge the substantial contributions
   of Josh Broch in helping to design, simulate, and implement the DSR
   protocol.  We thank him for his contributions to earlier versions of
   this document.

   We would also like to acknowledge the assistance of Robert V. Barron
   at Carnegie Mellon University.  Bob ported our DSR implementation
   from FreeBSD 2.2.7 into FreeBSD 3.3.

   Many valuable suggestions came from participants in the IETF process.
   We would particularly like to acknowledge Fred Baker, who provided
   extensive feedback on a previous version of this document, as well as
   the working group chairs, for their suggestions of previous versions
   of the document.



















Johnson, et al.               Experimental                    [Page 101]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Normative References

   [RFC791]       Postel, J., "Internet Protocol", STD 5, RFC 791,
                  September 1981.

   [RFC792]       Postel, J., "Internet Control Message Protocol", STD
                  5, RFC 792, September 1981.

   [RFC826]       Plummer, David C., "Ethernet Address Resolution
                  Protocol: Or converting network protocol addresses to
                  48.bit Ethernet address for transmission on Ethernet
                  hardware", STD 37, RFC 826, November 1982.

   [RFC1122]      Braden, R., "Requirements for Internet Hosts -
                  Communication Layers", STD 3, RFC 1122, October 1989.

   [RFC1700]      Reynolds, J. and J. Postel, "Assigned Numbers", STD 2,
                  RFC 1700, October 1994.  See also
                  http://www.iana.org/numbers.html.

   [RFC2003]      Perkins, C., "IP Encapsulation within IP", RFC 2003,
                  October 1996.  RFC 2003, October 1996.

   [RFC2119]      Bradner, S., "Key words for use in RFCs to Indicate
                  Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC2434]      Narten, T. and H. Alvestrand, "Guidelines for Writing
                  an IANA Considerations Section in RFCs", BCP 26, RFC
                  2434, October 1998.

Informative References

   [BANTZ94]      David F. Bantz and Frederic J. Bauchot.  Wireless LAN
                  Design Alternatives.  IEEE Network, 8(2):43-53,
                  March/April 1994.

   [BHARGHAVAN94] Vaduvur Bharghavan, Alan Demers, Scott Shenker, and
                  Lixia Zhang.  MACAW: A Media Access Protocol for
                  Wireless LAN's.  In Proceedings of the ACM SIGCOMM '94
                  Conference, pages 212-225. ACM, August 1994.

   [BROCH98]      Josh Broch, David A. Maltz, David B. Johnson, Yih-Chun
                  Hu, and Jorjeta Jetcheva.  A Performance Comparison of
                  Multi-Hop Wireless Ad Hoc Network Routing Protocols.
                  In Proceedings of the Fourth Annual ACM/IEEE
                  International Conference on Mobile Computing and
                  Networking, pages 85-97.  ACM/IEEE, October 1998.




Johnson, et al.               Experimental                    [Page 102]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   [CLARK88]      David D. Clark.  The Design Philosophy of the DARPA
                  Internet Protocols.  In Proceedings of the ACM SIGCOMM
                  '88 Conference, pages 106-114. ACM, August 1988.

   [FREEBSD]      The FreeBSD Project.  Project web page available at
                  http://www.freebsd.org/.

   [HU00]         Yih-Chun Hu and David B. Johnson.  Caching Strategies
                  in On-Demand Routing Protocols for Wireless Ad Hoc
                  Networks.  In Proceedings of the Sixth Annual ACM
                  International Conference on Mobile Computing and
                  Networking. ACM, August 2000.

   [HU01]         Yih-Chun Hu and David B. Johnson.  Implicit Source
                  Routing in On-Demand Ad Hoc Network Routing.  In
                  Proceedings of the Second Symposium on Mobile Ad Hoc
                  Networking and Computing (MobiHoc 2001), pages 1-10,
                  October 2001.

   [HU02]         Yih-Chun Hu, Adrian Perrig, and David B. Johnson.
                  Ariadne:  A Secure On-Demand Routing Protocol for Ad
                  Hoc Networks.  In Proceedings of the Eighth Annual
                  International Conference on Mobile Computing and
                  Networking (MobiCom 2002), pages 12-23, September
                  2002.

   [IEEE80211]    IEEE Computer Society LAN MAN Standards Committee.
                  Wireless LAN Medium Access Control (MAC) and Physical
                  Layer (PHY) Specifications, IEEE Std 802.11-1997.  The
                  Institute of Electrical and Electronics Engineers, New
                  York, New York, 1997.

   [JOHANSSON99]  Per Johansson, Tony Larsson, Nicklas Hedman, Bartosz
                  Mielczarek, and Mikael Degermark.  Scenario-based
                  Performance Analysis of Routing Protocols for Mobile
                  Ad-hoc Networks.  In Proceedings of the Fifth Annual
                  ACM/IEEE International Conference on Mobile Computing
                  and Networking, pages 195-206. ACM/IEEE, August 1999.

   [JOHNSON94]    David B. Johnson.  Routing in Ad Hoc Networks of
                  Mobile Hosts.  In Proceedings of the IEEE Workshop on
                  Mobile Computing Systems and Applications, pages 158-
                  163. IEEE Computer Society, December 1994.








Johnson, et al.               Experimental                    [Page 103]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   [JOHNSON96a]   David B. Johnson and David A. Maltz.  Dynamic Source
                  Routing in Ad Hoc Wireless Networks.  In Mobile
                  Computing, edited by Tomasz Imielinski and Hank Korth,
                  chapter 5, pages 153-181. Kluwer Academic Publishers,
                  1996.

   [JOHNSON96b]   David B. Johnson and David A. Maltz.  Protocols for
                  Adaptive Wireless and Mobile Networking.  IEEE
                  Personal Communications, 3(1):34-42, February 1996.

   [JUBIN87]      John Jubin and Janet D. Tornow.  The DARPA Packet
                  Radio Network Protocols.  Proceedings of the IEEE,
                  75(1):21-32, January 1987.

   [KARN90]       Phil Karn.  MACA---A New Channel Access Method for
                  Packet Radio.  In ARRL/CRRL Amateur Radio 9th Computer
                  Networking Conference, pages 134-140. American Radio
                  Relay League, September 1990.

   [LAUER95]      Gregory S. Lauer.  Packet-Radio Routing.  In Routing
                  in Communications Networks, edited by Martha E.
                  Steenstrup, chapter 11, pages 351-396. Prentice-Hall,
                  Englewood Cliffs, New Jersey, 1995.

   [MALTZ99a]     David A. Maltz, Josh Broch, Jorjeta Jetcheva, and
                  David B. Johnson.  The Effects of On-Demand Behavior
                  in Routing Protocols for Multi-Hop Wireless Ad Hoc
                  Networks.  IEEE Journal on Selected Areas of
                  Communications, 17(8):1439-1453, August 1999.

   [MALTZ99b]     David A. Maltz, Josh Broch, and David B. Johnson.
                  Experiences Designing and Building a Multi-Hop
                  Wireless Ad Hoc Network Testbed.  Technical Report
                  CMU-CS-99-116, School of Computer Science, Carnegie
                  Mellon University, Pittsburgh, Pennsylvania, March
                  1999.

   [MALTZ00]      David A. Maltz, Josh Broch, and David B. Johnson.
                  Quantitative Lessons From a Full-Scale Multi-Hop
                  Wireless Ad Hoc Network Testbed.  In Proceedings of
                  the IEEE Wireless Communications and Networking
                  Conference. IEEE, September 2000.

   [MALTZ01]      David A. Maltz, Josh Broch, and David B. Johnson.
                  Lessons From a Full-Scale MultiHop Wireless Ad Hoc
                  Network Testbed.  IEEE Personal Communications,
                  8(1):8-15, February 2001.




Johnson, et al.               Experimental                    [Page 104]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


   [MONARCH]      Rice University Monarch Project.  Monarch Project Home
                  Page.  Available at http://www.monarch.cs.rice.edu/.

   [NS-2]         The Network Simulator -- ns-2.  Project web page
                  available at http://www.isi.edu/nsnam/ns/.

   [PAPADIMITRATOS02]
                  Panagiotis Papadimitratos and Zygmunt J. Haas.  Secure
                  Routing for Mobile Ad Hoc Networks.  In SCS
                  Communication Networks and Distributed Systems
                  Modeling and Simulation Conference (CNDS 2002),
                  January 2002.

   [PERLMAN92]    Radia Perlman.  Interconnections:  Bridges and
                  Routers.  Addison-Wesley, Reading, Massachusetts,
                  1992.

   [RFC793]       Postel, J., "Transmission Control Protocol", STD 7,
                  RFC 793, September 1981.

   [RFC2131]      Droms, R., "Dynamic Host Configuration Protocol", RFC
                  2131, March 1997.

   [RFC2460]      Deering, S. and R. Hinden, "Internet Protocol, Version
                  6 (IPv6) Specification", RFC 2460, December 1998.

   [SONG01]       Alex Song.  picoNet II: A Wireless Ad Hoc Network for
                  Mobile Handheld Devices.  Submitted for the degree of
                  Bachelor of Engineering (Honours) in the division of
                  Electrical Engineering, Department of Information
                  Technology and Electrical Engineering, University of
                  Queensland, Australia, October 2001.  Available at
                  http://piconet.sourceforge.net/thesis/index.html.

   [TURNER90]     Paul Turner.  NetWare Communications Processes.
                  NetWare Application Notes, Novell Research, pages 25-
                  91, September 1990.

   [WRIGHT95]     Gary R. Wright and W. Richard Stevens.  TCP/IP
                  Illustrated, Volume 2:  The Implementation.  Addison-
                  Wesley, Reading, Massachusetts, 1995.










Johnson, et al.               Experimental                    [Page 105]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Authors' Addresses



   David B. Johnson
   Rice University
   Computer Science Department, MS 132
   6100 Main Street
   Houston, TX 77005-1892
   USA

   Phone: +1 713 348-3063
   Fax:   +1 713 348-5930
   EMail: dbj@cs.rice.edu


   David A. Maltz
   Microsoft Research
   One Microsoft Way
   Redmond, WA 98052
   USA

   Phone: +1 425 706-7785
   Fax:   +1 425 936-7329
   EMail: dmaltz@microsoft.com


   Yih-Chun Hu
   University of Illinois at Urbana-Champaign
   Coordinated Science Lab
   1308 West Main St, MC 228
   Urbana, IL 61801
   USA

   Phone: +1 217 333-4220
   EMail: yihchun@uiuc.edu

















Johnson, et al.               Experimental                    [Page 106]

RFC 4728          The Dynamic Source Routing Protocol      February 2007


Full Copyright Statement



   Copyright (C) The IETF Trust (2007).

   This document is subject to the rights, licenses and restrictions
   contained in BCP 78, and except as set forth therein, the authors
   retain all their rights.

   This document and the information contained herein are provided on an
   "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS
   OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY, THE IETF TRUST AND
   THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS
   OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF
   THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED
   WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE.

Intellectual Property



   The IETF takes no position regarding the validity or scope of any
   Intellectual Property Rights or other rights that might be claimed to
   pertain to the implementation or use of the technology described in
   this document or the extent to which any license under such rights
   might or might not be available; nor does it represent that it has
   made any independent effort to identify any such rights.  Information
   on the procedures with respect to rights in RFC documents can be
   found in BCP 78 and BCP 79.

   Copies of IPR disclosures made to the IETF Secretariat and any
   assurances of licenses to be made available, or the result of an
   attempt made to obtain a general license or permission for the use of
   such proprietary rights by implementers or users of this
   specification can be obtained from the IETF on-line IPR repository at
   http://www.ietf.org/ipr.

   The IETF invites any interested party to bring to its attention any
   copyrights, patents or patent applications, or other proprietary
   rights that may cover technology that may be required to implement
   this standard.  Please address the information to the IETF at
   ietf-ipr@ietf.org.

Acknowledgement



   Funding for the RFC Editor function is currently provided by the
   Internet Society.







Johnson, et al.               Experimental                    [Page 107]