RFC 2745






Network Working Group                                          A. Terzis
Request for Comments: 2745                                          UCLA
Category: Standards Track                                      B. Braden
                                                                     ISI
                                                              S. Vincent
                                                           Cisco Systems
                                                                L. Zhang
                                                                    UCLA
                                                            January 2000


                        RSVP Diagnostic Messages

Status of this Memo



   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

Copyright Notice



   Copyright (C) The Internet Society (2000).  All Rights Reserved.

Abstract



   This document specifies the RSVP diagnostic facility, which allows a
   user to collect information about the RSVP state along a path.  This
   specification describes the functionality, diagnostic message
   formats, and processing rules.

1.  Introduction



   In the basic RSVP protocol [RSVP], error messages are the only means
   for an end host to receive feedback regarding a failure in setting up
   either path state or reservation state.  An error message carries
   back only the information from the failed point, without any
   information about the state at other hops before or after the
   failure.  In the absence of failures, a host receives no feedback
   regarding the details of a reservation that has been put in place,
   such as whether, or where, or how, its own reservation request is
   being merged with that of others.  Such missing information can be
   highly desirable for debugging purposes, or for network resource
   management in general.






Terzis, et al.              Standards Track                     [Page 1]

RFC 2745                RSVP Diagnostic Messages            January 2000


   This document specifies the RSVP diagnostic facility, which is
   designed to fill this information gap.  The diagnostic facility can
   be used to collect and report RSVP state information along the path
   from a receiver to a specific sender.  It uses Diagnostic messages
   that are independent of other RSVP control messages and produce no
   side-effects; that is, they do not change any RSVP state at either
   nodes or hosts.  Similarly, they provide not an error report but
   rather a collection of requested RSVP state information.

   The RSVP diagnostic facility was designed with the following goals:

   -  To collect RSVP state information from every RSVP-capable hop
      along a path defined by path state, either for an existing
      reservation or before a reservation request is made.  More
      specifically, we want to be able to collect information about
      flowspecs, refresh timer values, and reservation merging at each
      hop along the path.

   -  To collect the IP hop count across each non-RSVP cloud.

   -  To avoid diagnostic packet implosion or explosion.

   The following is specifically identified as a non-goal:

   -  Checking the resource availability along a path.  Such
      functionality may be useful for future reservation requests, but
      it would require modifications to existing admission control
      modules that is beyond the scope of RSVP.

2.  Overview



   The diagnostic facility introduces two new RSVP message types:
   Diagnostic Request (DREQ) and Diagnostic Reply (DREP).  A DREQ
   message can be originated by a client in a "requester" host, which
   may or may not be a participant of the RSVP session to be diagnosed.
   A client in the requester host invokes the RSVP diagnostic facility
   by generating a DREQ packet and sending it towards the LAST-HOP node,
   which should be on the RSVP path to be diagnosed. This DREQ packet
   specifies the RSVP session and a sender host for that session.
   Starting from the LAST-HOP, the DREQ packet collects information
   hop-by-hop as it is forwarded towards the sender (see Figure 1),
   until it reaches the ending node.  Specifically, each RSVP-capable
   hop adds to the DREQ message a response (DIAG_RESPONSE) object
   containing local RSVP state for the specified RSVP session.







Terzis, et al.              Standards Track                     [Page 2]

RFC 2745                RSVP Diagnostic Messages            January 2000


   When the DREQ packet reaches the ending node, the message type is
   changed to Diagnostic Reply (DREP) and the completed response is sent
   to the original requester node.  Partial responses may also be
   returned before the DREQ packet reaches the ending node if an error
   condition along the path, such as "no path state", prevents further
   forwarding of the DREQ packet.  To avoid packet implosion or
   explosion, all diagnostic packets are forwarded via unicast only.

   Thus, there are generally three nodes (hosts and/or routers) involved
   in performing the diagnostic function: the requester node, the
   starting node, and the ending node, as shown in Figure 1.  It is
   possible that the client invoking the diagnosis function may reside
   directly on the starting node, in which case that the first two nodes
   are the same.  The starting node is named "LAST-HOP", meaning the
   last-hop of the path segment to be diagnosed.  The LAST-HOP node can
   be either a receiver node or an intermediate node along the path.
   The ending node is usually the specified sender host.  However, the
   client can limit the length of the path segment to be diagnosed by
   specifying a hop-count limit in the DREQ message.


                  LAST-HOP                  Ending
     Receiver        node                     node           Sender
         __           __         __            __              __
        |  |---------|  |------>|  |--> ...-->|  |--> ...---->|  |
        |__|         |__| DREQ  |__|   DREQ   |__|   DREQ     |__|
                      ^                         .              |
                      |                         .              |
                      | DREQ                    . DREP         | DREP
                      |                         .              |
                     _|_               DREP     V              V
        Requester   |   | <------------------------------------
        (client)    |___|

                         Figure 1


   DREP packets can be unicast from the ending node back to the
   requester either directly or hop-by-hop along the reverse of the path
   taken by the DREQ message to the LAST-HOP, and thence to the
   requester.  The direct return is faster and more efficient, but the
   hop-by-hop reverse-path route may be the only choice if the packets
   have to cross firewalls.  Hop-by-hop return is accomplished using an
   optional ROUTE object, which is built incrementally to contain a list
   of node addresses that the DREQ packet has passed through.  The ROUTE
   object is then used in reverse as a source route to forward the DREP
   hop-by-hop back to the LAST-HOP node.




Terzis, et al.              Standards Track                     [Page 3]

RFC 2745                RSVP Diagnostic Messages            January 2000


   A DREQ message always consists of a single unfragmented IP datagram.
   On the other hand, one DREQ message can generate multiple DREP
   packets, each containing a fragment of the total DREQ message.  When
   the path consists of many hops, the total length of a DREP message
   will exceed the MTU size before reaching the ending node; thus, the
   message has to be fragmented.  Relying on IP fragmentation and
   reassembly, however, can be problematic, especially when DREP
   messages are returned to the requester hop-by-hop, in which case
   fragmentation/reassembly would have to be performed at every hop.  To
   avoid such excessive overhead, we let the requester define a default
   path MTU size that is carried in every DREQ packet.  If an
   intermediate node finds that the default MTU size is bigger than the
   MTU of the incoming interface, it reduces the default MTU size to the
   MTU size of the incoming interface. If an intermediate node detects
   that a DREQ packet size is larger than the default MTU size, it
   returns to the requester (in either manner described above) a DREP
   fragment containing accumulated responses.  It then removes these
   responses from the DREQ and continues to forward it.  The requester
   node can reassemble the resulting DREP fragments into a complete DREP
   message.

   When discussing diagnostic packet handling, this document uses
   direction terminology that is consistent with the RSVP functional
   specification [RSVP], relative to the direction of data packet flow.
   Thus, a DREQ packet enters a node through an "outgoing interface" and
   is forwarded towards the sender through an "incoming interface",
   because DREQ packets travel in the reverse direction to the data
   flow.

   Notice that DREQ packets can be forwarded only after the RSVP path
   state has been set up.  If no path state exists, one may resort to
   the traceroute or mtrace facility to examine whether the
   unicast/multicast routing is working correctly.

3.  Diagnostic Packet Format



   Like other RSVP messages, DREQ and DREP messages consist of an RSVP
   Common Header followed by a variable set of typed RSVP data objects.
   The following sequence must be used:












Terzis, et al.              Standards Track                     [Page 4]

RFC 2745                RSVP Diagnostic Messages            January 2000


           +-----------------------------------+
           |        RSVP Common Header         |
           +-----------------------------------+
           |         Session object            |
           +-----------------------------------+
           |      Next-Hop RSVP_HOP object     |
           +-----------------------------------+
           |       DIAGNOSTIC object           |
           +-----------------------------------+
           |    (optional) DIAG_SELECT object  |
           +-----------------------------------+
           |    (optional) ROUTE object        |
           +-----------------------------------+
           | zero or more DIAG_RESPONSE objects|
           +-----------------------------------+

   The session object identifies the RSVP session for which the state
   information is being collected.  We describe each of the other parts.

3.1.  RSVP Message Common Header



   The RSVP message common header is defined in [RSVP].  The following
   specific exceptions and extensions are needed for DREP and DREQ.

   Type field: define:

          Type = 8: DREQ     Diagnostic Request

          Type = 9: DREP     Diagnostic Reply

   RSVP length:

      If this is a DREP message and the MF flag in the DIAGNOSTIC object
      (see below) is set, this field indicates the length of this single
      DREP fragment rather than the total length of the complete DREP
      reply message (which cannot generally be known in advance).

3.2.  Next-Hop RSVP_HOP Object



   This RSVP_HOP object carries the LIH of the interface through which
   the DREQ should be received at the upstream node. This object is
   updated hop-by hop. It is used for the same reasons that a RESV
   message contains an RSVP_HOP object: to distinguish logical
   interfaces and avoid problems caused by routing asymmetries and non-
   RSVP clouds.






Terzis, et al.              Standards Track                     [Page 5]

RFC 2745                RSVP Diagnostic Messages            January 2000


   While the IP address is not really used during DREQ processing, for
   consistency with the use of the RSVP_HOP object in other RSVP
   messages, the IP address in the RSVP_HOP object to contain the
   address of the interface through which the DREQ was sent.

3.3.  DIAGNOSTIC Object



   A DIAGNOSTIC object contains the common diagnostic control
   information in both DREQ and DREP messages.

   o IPv4 DIAGNOSTIC object: Class = 30, C-Type = 1

    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    | Max-RSVP-hops | RSVP-hop-count|         Reserved            |MF|
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                          Request ID                           |
    +---------------+---------------+---------------+---------------+
    |           Path MTU            |     Fragment Offset           |
    +---------------+---------------+---------------+---------------+
    |                         LAST-HOP Address                      |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                                                               |
    |                     SENDER_TEMPLATE object                    |
    |                                                               |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                                                               |
    |                 Requester FILTER_SPEC object                  |
    |                                                               |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   Here all IP addresses use the 4 byte IPv4 format, both explicitly in
   the LAST-HOP Address and by using the IPv4 forms of the embedded
   FILTER_SPEC and RSVP_HOP objects.

   o IPv6 DIAGNOSTIC object: Class = 30, C-Type = 2

   The format is the same, except all explicit and embedded IP addresses
   are 16 byte IPv6 addresses.

   The fields are as follows:

   Max-RSVP-hops

      An octet specifying the maximum number of RSVP hops over which
      information will be collected.  If an error condition in the
      middle of the path prevents the DREQ packet from reaching the
      specified ending node, the Max-RSVP-hops field may be used to
      perform an expanding-length search to reach the point just before



Terzis, et al.              Standards Track                     [Page 6]

RFC 2745                RSVP Diagnostic Messages            January 2000


      the problem.  If this value is 1, the starting node and the ending
      node of the query will be the same.  If it is zero, there is no
      hop limit.

   RSVP-hop-count

      Records the number of RSVP hops that have been traversed so far.
      If the starting and ending nodes are the same, this value will be
      1 in the resulting DREP message.

   Fragment Offset

      Indicates where this DREP fragment belongs in the complete DREP
      message, measured in octets.  The first fragment has offset zero.
      Fragment Offset is used also to determine if a DREQ message
      containing zero DIAG_RESPONSE objects should be processed at an
      RSVP capable node.

   MF flag

      Flag means "more fragments".  It must be set to zero (0) in all
      DREQ messages.  It must be set to one (1) in all DREP packets that
      carry partial results and are returned by intermediate nodes due
      to the MTU limit.  When the DREQ message is converted to a DREP
      message in the ending node, the MF flag must remain zero.

   Request ID

      Identifies an individual DREQ message and the corresponding DREP
      message (or all the fragments of the reply message).

      One possible way to define the Request ID would use 16 bits to
      specify the ID of the process making the query and 16 bits to
      distinguish different queries from this process.

   Path MTU

      Specifies a default MTU size in octets for DREP and DREQ messages.
      This value should not be smaller than the size of the "base" DREQ
      packet. A "base" DREQ packet is one that contains a Common Header,
      a Session object, a Next-Hop RSVP_HOP object, a DIAGNOSTIC object,
      an empty ROUTE object and a single default DIAG_RESPONSE (see
      below).  The assumption made here is that a diagnostic packet of
      this size can always be forwarded without IP fragmentation.







Terzis, et al.              Standards Track                     [Page 7]

RFC 2745                RSVP Diagnostic Messages            January 2000


   LAST-HOP Address

      The IP address of the LAST-HOP node.  The DREQ message starts
      collecting information at this node and proceeds toward the
      sender.

   SENDER_TEMPLATE object

      This IPv4/IPv6 SENDER_TEMPLATE object contains the IP address and
      the port of a sender for the session being diagnosed.  The DREQ
      packet is forwarded hop-by-hop towards this address.

   Requester FILTER_SPEC Object

      This IPv4/IPv6 FILTER_SPEC object contains the IP address and the
      port from which the request originated and to which the DREP
      message(s) should be sent.

3.4.  DIAG_SELECT Object



   o DIAG_SELECT Class = 33, C-Type = 1.

   A Diagnostic message may optionally contain a DIAG_SELECT object to
   specify which specific RSVP objects should be reported in a
   DIAG_RESPONSE object.  In the absence of a DIAG_SELECT object, the
   DIAG_RESPONSE object added by the node will contain a default set of
   object types (see DIAG_RESPONSE object below).

   The DIAG_SELECT object contains a list of [Class, C-type] pairs, in
   the following format:

    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |    class      |     C-Type    |    class      |     C-Type    |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    //                                                             //
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |    class      |     C-Type    |    class      |     C-Type    |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   When a DIAG_SELECT object is included in a DREQ message, each RSVP
   node along the path will add a DIAG_RESPONSE object containing
   response objects (see below) whose classes and C-Types match entries
   in the DIAG_SELECT list (and are from matching path and reservation
   state). A C-type octet of zero is a 'wildcard', matching any C-Type
   associated with the associated class.






Terzis, et al.              Standards Track                     [Page 8]

RFC 2745                RSVP Diagnostic Messages            January 2000


   Depending on the type of objects requested, a node can find the
   associated information in the path or reservation state stored for
   the session described in the SESSION object. Specifically,
   information for the RSVP_HOP,SENDER_TEMPLATE, SENDER_TSPEC, ADSPEC
   objects can be extracted from the node's path state, while
   information for the FLOWSPEC, FILTER_SPEC, CONFIRM, STYLE and SCOPE
   objects can be found in the node's reservation state (if existent).

   If the number of [Class, C-Type] pairs is odd, the last two octets of
   the DIAG_SELECT object must be  zero. A maximum DIAG_SELECT object is
   one that contains the [Class, C-type] pairs for all the RSVP objects
   that can be requested in a Diagnostic query.

3.5.  ROUTE Object



   A diagnostic message may contain a ROUTE object, which is used to
   record the route of the DREQ message and as a source route for
   returning the DREP message(s) hop-by-hop.

   o IPv4 ROUTE object: Class = 31, C-Type = 1.

    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |             reserved                          |    R-pointer  |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                                                               |
    +                     RSVP Node List                            |
    |                                                               |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   This message signifies how the reply should be returned.  If it does
   not exist in the DREQ packet then DREP packets should be sent to the
   requester directly. If it does exist, DREP packets must be returned
   hop-by-hop along the reverse path to the LAST-HOP node and thence to
   the requester node.

   An empty ROUTE object is one that has an empty RSVP Node list and R-
   pointer is equal to zero.

   RSVP Node List

      A list of RSVP node IPv4 addresses.  The number of addresses in
      this list can be computed from the object size.

   R-pointer

      Used in DREP messages only (see Section 4.2 for details), but it
      is incremented as each hop adds its incoming interface address in
      the ROUTE object.



Terzis, et al.              Standards Track                     [Page 9]

RFC 2745                RSVP Diagnostic Messages            January 2000


   o IPv6 ROUTE object: Class = 31, C-Type = 2

   The same, except RSVP Node List contains IPv6 addresses.

   In a DREQ message, RSVP Node List specifies all RSVP hops between the
   LAST-HOP address specified in the DIAGNOSTIC object, and the last
   RSVP node the DREQ message has visited.  In a DREP message, RSVP Node
   List specifies all RSVP hops between the LAST-HOP and the node that
   returns this DREP message.

3.6.  DIAG_RESPONSE Object



   Each RSVP node attaches a DIAG_RESPONSE object to each DREQ message
   it receives, before forwarding the message.  The DIAG_RESPONSE object
   contains the state to be reported for this node.  It has a fixed-
   format header and then a variable list of RSVP state objects, or
   "response objects".

   o IPv4 DIAG_RESPONSE object: Class = 32, C-Type = 1.

    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                       DREQ Arrival Time                       |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                  Incoming Interface Address                   |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                  Outgoing Interface Address                   |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                 Previous-RSVP-Hop Router Address              |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |   D-TTL       |M|R-err|  K    |      Timer value              |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                                                               |
    |                  (optional) TUNNEL object                     |
    |                                                               |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
    |                                                               |
    //                       Response objects                      //
    |                                                               |
    +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+


   o IPv6 DIAG_RESPONSE object: Class = 32, C-Type = 2.

   This object has the same format, except that all explicit and
   embedded IP addresses are IPv6 addresses.

   The fields are as follows:




Terzis, et al.              Standards Track                    [Page 10]

RFC 2745                RSVP Diagnostic Messages            January 2000


   DREQ Arrival Time

      A 32-bit NTP timestamp specifying the time the DREQ message
      arrived at this node.  The 32-bit form of an NTP timestamp
      consists of the middle 32 bits of the full 64-bit form, that is,
      the low 16 bits of the integer part and the high 16 bits of the
      fractional part.

   Incoming Interface Address

      Specifies the IP address of the interface on which messages from
      the sender are expected to arrive, or 0 if unknown.

   Outgoing Interface Address

      Specifies the IP address of the interface through which the DREQ
      message arrived and to which messages from the given sender and
      for the specified session address flow, or 0 if unknown.

   Previous-RSVP-Hop Router Address

      Specifies the IP address from which this node receives RSVP PATH
      messages for this source, or 0 if unknown.  This is also the
      interface to which the DREQ will be forwarded.

   D-TTL

      The number of IP hops this DREQ message traveled from the down-
      stream RSVP node to the current node.

   M flag

      A single-bit flag which indicates whether the reservation
      described by the response objects is merged with reservations from
      other down-stream interfaces when being forwarded upstream.

   R-error

      A 3-bit field that indicates error conditions at a node. Currently
      defined values are:

           0x00: no error
           0x01: No PATH state
           0x02: packet too big
           0x04: ROUTE object too big






Terzis, et al.              Standards Track                    [Page 11]

RFC 2745                RSVP Diagnostic Messages            January 2000


   K

      The refresh timer multiple (defined in [RSVP]).

   Timer value

      The local refresh timer value in seconds.

   The set of response objects to be included at the end of the
   DIAG_RESPONSE object is determined by a DIAG_SELECT object, if one is
   present.  If no DIAG_SELECT object is present, the response objects
   belong to the default list of classes:

      SENDER_TSPEC object      FILTER_SPEC object      FLOWSPEC object
      STYLE object

   Any C-Type present in the local RSVP state will be used.  These
   response objects may be in any order but they must all be at the end
   of the DIAG_RESPONSE object.

   A default DIAG_RESPONSE object is one containing the default list of
   classes described above.

3.7.  TUNNEL Object



   The optional TUNNEL object should be inserted when a DREQ message
   arrives at an RSVP node that acts as a tunnel exit point.

   The TUNNEL object provides the mapping between the end-to-end RSVP
   session that is being diagnosed and the RSVP session over the tunnel.
   This mapping information allows the diagnosis client to conduct
   diagnosis over the involved tunnel session, by invoking a separate
   Diagnostic query for the corresponding Tunnel Session and Tunnel
   Sender.  Keep in mind, however, that multiple end-to-end sessions may
   all map to one pre-configured tunnel session that may have totally
   different parameter settings.

   The tunnel object is defined in the RSVP Tunnel Specification
   [RSVPTUN].

4.  Diagnostic Packet Forwarding Rules



4.1.  DREQ Packet Forwarding



   DREQ messages are forwarded  hop-by-hop via unicast from the LAST-HOP
   address to the Sender address, as specified in the DIAGNOSTIC object.
   If an RSVP capable node, other than the LAST-HOP node, receives a
   DREQ message  that contains no DIAG_RESPONSE objects and has a zero



Terzis, et al.              Standards Track                    [Page 12]

RFC 2745                RSVP Diagnostic Messages            January 2000


   Fragment Offset, the node should forward the DREQ packet towards the
   LAST-HOP without doing any of the processing mentioned below. The
   reason is that such conditions apply only for nodes downstream of the
   LAST-HOP where no information should be collected.

   Processing begins when a DREQ message, DREQ_in, arrives at a node.

       1. Create a new DIAG_RESPONSE object. Compute the IP hop count
          from the previous RSVP hop. This is done by subtracting the
          value of the TTL value in the IP header from Send_TTL in the
          RSVP common header.  Save the result in the D-TTL field of the
          DIAG_RESPONSE object.

       2. Set the DREQ Arrival Time and the Outgoing Interface Address
          in the DIAG_RESPONSE object.  If this node is the LAST-HOP,
          then the Out- going Interface Address field in the
          DIAG_RESPONSE object contains the following value depending on
          the session being diagnosed.

         *  If the session in question is a unicast session, then the
            Out-going Interface Address field contains the address of
            the interface LAST-HOP uses to send PATH messages and data
            to the receiver specified by the session address.

         *  Otherwise, if it is a multicast session and there is at
            least one receiver for this session, LAST_HOP should use the
            address of one of local interfaces used to reach one of the
            receivers.

         *  Otherwise Outgoing Interface Address should be zero.

       3. Increment the RSVP-hop-count field in the DIAGNOSTIC message
          object by one.

       4. If no PATH state exists for the specified session, set R-error
          = 0x01 (No PATH state) and goto step 7.

       5. Set the rest of the fields in the DIAG_RESPONSE object. If
          DREQ_in contains a DIAG_SELECT object, the response object
          classes are those specified in the DIAG_SELECT; otherwise,
          they are SENDER_TSPEC, STYLE, and FLOWSPEC objects. If no
          reservation state exists for the specified RSVP session, the
          DIAG_RESPONSE object will contain no FLOWSPEC, FILTER_SPEC or
          STYLE object. If neither PATH nor reservation state exists for
          the specified RSVP session, then no response objects will be
          appended to the DIAG_RESPONSE object.





Terzis, et al.              Standards Track                    [Page 13]

RFC 2745                RSVP Diagnostic Messages            January 2000


       6. If RSVP-hop-count is less than Max-RSVP-hops and this node is
          not the sender, then the DREQ is eligible for forwarding; set
          the Path MTU to the min of the Path MTU and the MTU size of
          the incoming interface for the sender being diagnosed.

       7. If the size of DREQ_in plus the size of the new DIAG_RESPONSE
          object plus the size of an IP address (if a ROUTE object
          exists and R-error= 0) is larger than Path MTU, then the new
          diagnostic message will be too large to be forwarded or
          returned without fragmentation; set the "packet too big"
          (0x02) error bit in DIAG_RESPONSE and goto Step SD1 in
          Send_DREP (below).

       8. If the "No PATH state" (0x01) error bit is set or if RSVP-
          hop-count is equal to Max-RSVP-hops or if this node is the
          sender, then the DREQ cannot be forwarded further; goto Step
          10.

       9. Forward the DREQ towards the sender, as follows.  If a ROUTE
          object exists, append the "Incoming Interface Address" to the
          end of the ROUTE object and increment R-Pointer by one.
          Update the Next-Hop RSVP_HOP object, append the new
          DIAG_RESPONSE object to the list of DIAG_RESPONSE object, and
          update the message length field in the RSVP common header
          accordingly. Finally, recompute the checksum, forward DREQ_in
          to the next hop towards the sender, and return.

      10. Turn the DREQ into a DREP and return to the requester, as
          follows.  Append the DIAG_RESPONSE object to the end of
          DREQ_in and update the packet length.  If a ROUTE object is
          present in the message, decrement the R-pointer and set target
          address to the last address in the ROUTE object, otherwise set
          target address to the requester address. Change the Type Field
          in the Common header from DREQ to DREP.  Finally, recompute
          the checksum, send the DREP to the target address, and return.
          Note that the MF bit must be off in this case.

   Send_DREP:

   This sequence is entered if the DREQ message augmented with the new
   DIAG_RESPONSE object is too large to be forwarded towards the sender
   or, if it is not eligible for forwarding, too large to be returned as
   a DREP.

   SD1. Make a copy of DREQ_in and change the message type field from
        DREQ to DREP.  Trim all DIAG_RESPONSE objects from DREQ_in and
        adjust the Fragment Offset.  The DREP message contains the
        DIAG_RESPONSE objects accumulated by prior nodes.



Terzis, et al.              Standards Track                    [Page 14]

RFC 2745                RSVP Diagnostic Messages            January 2000


   SD2. Send the DREP message towards the requester, as follows.  If a
        ROUTE object is present in the DREP message, decrement the R-
        pointer and set target address to the last address in the ROUTE
        object, otherwise set target address to the requester address.
        Set the MF bit, recompute the checksum and send the DREP message
        back to the target address.

   SD3. If the reduced size of DREQ_in plus the size of DIAG_RESPONSE
        plus the size of an IP address (if a ROUTE object exists) is
        smaller than or equal to Path MTU, then return to Step 8 of the
        main DREQ processing sequence above.

   SD4. If a ROUTE object exists, replace the ROUTE object in DREQ_in
        with an empty ROUTE object and turn on the "ROUTE object too
        big" (0x04) error bit in the DIAG_RESPONSE.  In either case,
        return to Step 8 of the main DREQ processing sequence above.

4.2.  DREP Forwarding



   When a ROUTE object is present, DREP messages are forwarded hop-by-
   hop towards the requester, by reversing the route as listed in the
   ROUTE object. Otherwise, DREP messages are sent directly to the
   original requester.

   When a node receives a DREP message, it simply decreases R-pointer by
   one (address length), recomputes the checksum and forwards the
   message to the address pointed to by R-pointer in the route list. If
   a node, other than the LAST-HOP, receives a DREP packet where R-
   pointer is equal to zero, it must send it directly to the requester.

   When the LAST-HOP node receives a DREP message, it sends the message
   to the requester.

4.3.  MTU Selection and Adjustment



   Because the DREQ message carries the allowed MTU size of previous
   hops that the DREP messages will later traverse, this unique feature
   allows easy semantic fragmentation as described above.  Whenever the
   DREQ message approaches the size of Path MTU, it can be trimmed
   before being forwarded again.

   When a requester sends a DREQ message, the Path MTU field in the
   DIAGNOSTIC object can be set to a configured default value. It is
   possible that the original Path MTU value is chosen larger than the
   actual MTU value along some portion of the path being traced.
   Therefore each intermediate RSVP node must check the MTU value when
   processing a DREQ message.  If the specified MTU value is larger than




Terzis, et al.              Standards Track                    [Page 15]

RFC 2745                RSVP Diagnostic Messages            January 2000


   the MTU of the incoming interface (that the DREQ message will be
   forwarded to), the node changes the MTU value in the header to the
   smaller value.

   Whenever a DREQ message size becomes larger than the Path MTU value,
   an intermediate RSVP node makes a copy of the message, converts it to
   a DREP message to send back, and then trims off the partial results
   from the DREQ message. If in this case also the DREQ cannot be
   forwarded upstream due to a large ROUTE object, the "ROUTE object too
   big" is set and the ROUTE object is trimmed. As a result of the ROUTE
   object trimming, DREP(s) will come hop-by-hop up to this node and
   will then immediately be forwarded to the requester address.

   Even if the steps shown above are followed there are a few cases
   where fragmentation at the IP layer will happen. For example, non-
   RSVP hops with smaller MTUs may exist before LAST-HOP is reached, or
   if the response is sent directly back to requester (as opposed to hop
   by hop) the DREP may take a different route to the requester than the
   DREQ took from the requester. Another case is when there exists a
   link with MTU smaller than the minimum Path MTU value defined in
   Section 3.3.

4.4.  Errors



   If an error condition prevents a DREP message from being forwarded
   further, the message is simply dropped.

   If an error condition, such as lack of PATH state, prevents a DREQ
   message from being forwarded further, the node must change the
   current message to DREP type and return it to the response address.

5.  Problem Diagnosis by Using RSVP Diagnostic Facility



5.1.  Across Firewalls



   Firewalls may cause problems in diagnostic message forwarding.  Let
   us look at two different cases.

   First, let us assume that the querier resides on a receiving host of
   the session to be examined.  In this case, firewalls should not
   prevent the forwarding of the diagnostic messages in a hop-by-hop
   manner, assuming that proper holes have been punched on the firewall
   to allow hop-by-hop forwarding of other RSVP messages.  The querier
   may start by not including a ROUTE object, which can give a faster
   response delivery and reduced overhead at intermediate nodes.
   However if no response is received, the querier may resend the DREQ
   message with a ROUTE object, specifying that a hop-by-hop reply
   should be sent.



Terzis, et al.              Standards Track                    [Page 16]

RFC 2745                RSVP Diagnostic Messages            January 2000


   If the requester is a third party host and is separated from the
   LAST-HOP address by a firewall (either the requester is behind a
   firewall, or the LAST-HOP is a node behind a firewall, or both), at
   this time we do not know any other solution but to change the LAST-
   HOP to a node that is on the same side of the firewall as the
   requester.

5.2.  Examination of RSVP Timers



   One can easily collect information about the current timer value at
   each RSVP hop along the way.  This will be very helpful in situations
   when the reservation state goes up and down frequently, to find out
   whether the state changes are due to improper setting of timer
   values, or K values (when across lossy links), or frequent routing
   changes.

5.3.  Discovering Non-RSVP Clouds



   The D-TTL field in each DIAG_RESPONSE object shows the number of
   routing hops between adjacent RSVP nodes.  Therefore any value
   greater than one indicates a non-RSVP cloud in between.  Together
   with the arrival timestamps (assuming NTP works), this value can also
   give some vague, though not necessarily accurate, indication of how
   big that cloud might be.  One might also find out all the
   intermediate non-RSVP nodes by running either unicast or multicast
   trace route.

5.4.  Discovering Reservation Merges



   The flowspec value in a DIAG_RESPONSE object specifies the amount of
   resources being reserved for the data stream defined by the filter
   spec in the same data block.  When this value of adjacent
   DIAG_RESPONSE objects differs, that is, a downstream node Rd has a
   smaller value than its immediate upstream node Ru, it indicates a
   merge of reservation with RSVP request(s) from other down stream
   interface(s) at Rd.  Further, in case of SE style reservation, one
   can examine how the different SE scopes get merged at each hop.

   In particular, if a receiver sends a DREQ message before sending its
   own reservation, it can discover (1) how many RSVP hops there are
   along the path between the specified sender and itself, (2) how many
   of the hops already have some reservation by other receivers, and (3)
   possibly a rough prediction of how its reservation request might get
   merged with other existing ones.







Terzis, et al.              Standards Track                    [Page 17]

RFC 2745                RSVP Diagnostic Messages            January 2000


5.5.  Error Diagnosis



   In addition to examining the state of a working reservation, RSVP
   diagnostic messages are more likely to be invoked when things are not
   working correctly.  For example, a receiver has reserved an adequate
   pipe for a specified incoming data stream, yet the observed delay or
   loss ratio is much higher than expected.  In this case the receiver
   can use the diagnostic facility to examine the reservation state at
   each RSVP hop along the way to find out whether the RSVP state is set
   up correctly, whether there is any black-hole along the way that
   caused RSVP message losses, or whether there are non-RSVP clouds, and
   where they are, that may have caused the performance problem.

5.6.  Crossing "Legacy" RSVP Routers



   Since this diagnosis facility was developed and added to RSVP after a
   number of RSVP implementations were in place, it is possible, or even
   likely, that when performing RSVP diagnosis, one may encounter one or
   more RSVP-capable nodes that do not understand diagnostic messages
   and drop them.  When this happens, the invoking client will get no
   response from its requests.

   One way to by-pass such "legacy" RSVP nodes is to perform RSVP
   diagnosis repeatedly, guided by information from traceroute, or
   mtrace in case of multicast.  When an RSVP diagnostic query times out
   (see next section), one may first use traceroute to get the list of
   nodes along the path, and then gradually increase the value of Max-
   RSVP-hops field in the DREQ message, starting from a low value until
   one no longer receives a response.  One can then try RSVP diagnosis
   again by starting with the first node (which is further upstream
   towards the sender) after the unresponding one.

   There are two problem with the method mentioned above in the case of
   unicast sessions. Both problems are related to the fact that
   traceroute information provides the path from the requester to the
   sender. The first problem is that the LAST-HOP may not be on the path
   from the requester to the sender. In this case we can get information
   only from the portion of the path from the LAST-HOP to the sender
   which intersects with the path from the requester to the sender. If
   routers that are not on the intersection of the two paths don't have
   PATH state for the session being diagnosed then they will reply with
   R-error=0x01. The requester can overcome this problem by sending a
   DREQ to every router on the path (from itself to the sender) until it
   reaches the first router that belongs to the path from the sender to
   the LAST-HOP.






Terzis, et al.              Standards Track                    [Page 18]

RFC 2745                RSVP Diagnostic Messages            January 2000


   The second problem is that traceroute provides the path from the
   requester to the sender which, due to routing asymmetries, may be
   different than the path traffic from the sender to the LAST-HOP uses.
   There is (at least) one case where this asymmetry will cause the
   diagnosis to fail. We present this case below.

                                Downstream Path                Sender
                                __         __            __       __
   Receiver             +------|  |<------|  |<-- ...---|  |-----|  |
      __          __   /       |__|       |__|          |__|     |__|
     |  |--....--|X |_/                    ^
     |__|        |__| \     Router B       |
                Black  \        __         |
                Hole    +----->|  |---->---+
                               |__| Upstream Path

                             Router A

                             Figure 2

   Here the first hop upstream of the black hole is different on the
   upstream path and the downstream path. Traceroute will indicate
   router A as the previous hop (instead of router B which is the right
   one). Sending a DREQ to router A will result in A responding with R-
   error 0x01 (No PATH State). If the two paths converge again then the
   requester can use the solution proposed above to get any (partial)
   information from the rest of the path.

   We don't have, for the moment, any complete solutions for the
   problematic scenarios described here.

6.  Comments on Diagnostic Client Implementation.



   Following the design principle that nodes in the network should not
   hold more than necessary state,  RSVP nodes are responsible only for
   forwarding Diagnostic messages and filling DIAG_RESPONSE objects.
   Additional diagnostic functionality should be carried out by the
   diagnostic clients.  Furthermore, if the diagnostic function is
   invoked from a third-party host, we should not require that host be
   running an RSVP daemon to perform the function.  Below we sketch out
   the basic functions that a diagnostic client daemon should carry out.

      1. Take input from the user about the session to be diagnosed, the
         last-hop and the sender address, the Max-RSVP-hops, and
         possibly the DIAG_SELECT list, create a DREQ message and send
         to the LAST-HOP RSVP node using raw IP message with protocol
         number 46 (RSVP).  If the user specified that the response
         should be sent hop-by-hop include an empty ROUTE object to the



Terzis, et al.              Standards Track                    [Page 19]

RFC 2745                RSVP Diagnostic Messages            January 2000


         DREQ message sent. Set the Path_MTU to the smaller of the user
         request and the MTU of the link through which the DREQ will be
         sent.

         The port of the UDP socket on which the Diagnostic Client is
         listening for replies should be included in the Requester
         FILTER_SPEC object.

      2. Set a retransmission timer, waiting for the reply (one or more
         DREP messages).  Listen to the specified UDP port for responses
         from the LAST-HOP RSVP node.

         The LAST-HOP RSVP node, upon receiving DREP messages, sends
         them to the Diagnostic Client as UDP packets, using the port
         supplied in the Requester FILTER_SPEC object.

      3. Upon receiving a DREP message to an outstanding diagnostic
         request, the client should clear the retransmission timer,
         check to see if the reply contains the complete result of the
         requested diagnosis.  If so, it should pass the result up to
         the invoking entity immediately.

      4. Reassemble DREP fragments.  If the first reply to an
         outstanding diagnostic request contains only a fragment of the
         expected result, the client should set up a reassembly timer in
         a way similar to IP packet reassembly timer.  If the timer goes
         off before all fragments arrive, the client should pass the
         partial result to the invoking entity.

      5. Use retransmission and reassembly timers to gracefully handle
         packet losses and reply fragment scenarios.

         In the absence of response to the first diagnostic request, a
         client should retransmit the request a few times.  If all the
         retransmissions also fail, the client should invoke traceroute
         or mtrace to obtain the list of hops along the path segment to
         be diagnosed, and then perform an iteration of diagnosis with
         increasing hop count as suggested in Section 5.6 in order to
         cross RSVP-capable but diagnosis-incapable nodes.

      6. If all the above efforts fail, the client must notify the
         invoking entity.









Terzis, et al.              Standards Track                    [Page 20]

RFC 2745                RSVP Diagnostic Messages            January 2000


7.  Security Considerations



   RSVP Diagnostics, as any other diagnostic tool, can be a security
   threat since it can reveal possibly sensitive RSVP state information
   to unwanted third parties.

   We feel that the threat is minimal, since as explained in the
   Introduction Diagnostics messages produce no side-effects and
   therefore they cannot change RSVP state in the nodes. In this respect
   RSVP Diagnostics is less a security threat than other diagnostic
   tools and protocols such as SNMP.

   Furthermore, processing of Diagnostic messages can be disabled if it
   is felt that is a security threat.

8.  Acknowledgments



   The idea of developing a diagnostic facility for RSVP was first
   suggested by Mark Handley of ACIRI.  Many thanks to Lee Breslau of
   AT&T Labs and John Krawczyk of Nortel Networks for their valuable
   comments on the first draft of this memo.  Lee Breslau, Bob Braden,
   and John Krawczyk contributed further comments after March 1996 IETF.
   Steven Berson provided valuable comments on various drafts of the
   memo. Tim Gleeson contributed an extensive list of editorial
   comments. We would also like to acknowledge Intel for providing a
   research grant as a partial support for this work. Subramaniam
   Vincent did most of this work while a graduate research assistant at
   the USC Information Sciences Institute (ISI).

9.  References



   [RSVP]    Braden, R., Zhang, L., Berson, S., Herzog, S. and S. Jamin,
             "Resource ReserVation Protocol -- Version 1 Functional
             Specification", RFC 2205, September 1997.

   [RSVPTUN] Terzis, A., Krawczyk, J., Wroclawski, J. and L. Zhang,
             "RSVP Operation Over IP Tunnels", RFC 2746, January 2000.














Terzis, et al.              Standards Track                    [Page 21]

RFC 2745                RSVP Diagnostic Messages            January 2000


10.  Authors' Addresses



   Andreas Terzis
   UCLA
   4677 Boelter Hall
   Los Angeles, CA 90095

   Phone:    310-267-2190
   EMail:    terzis@cs.ucla.edu


   Bob Braden
   USC Information Sciences Institute
   4676 Admiralty Way
   Marina del Rey, CA 90292

   Phone:    310 822-1511
   EMail:    braden@isi.edu


   Subramaniam Vincent
   Cisco Systems
   275, E Tasman Drive, MS SJC04/2/1
   San Jose, CA 95134

   Phone:    408 525 3474
   EMail:    svincent@cisco.com


   Lixia Zhang
   UCLA
   4531G Boelter Hall
   Los Angeles, CA  90095

   Phone:    310-825-2695
   EMail:    lixia@cs.ucla.edu















Terzis, et al.              Standards Track                    [Page 22]

RFC 2745                RSVP Diagnostic Messages            January 2000


10.  Full Copyright Statement

   Copyright (C) The Internet Society (2000).  All Rights Reserved.

   This document and translations of it may be copied and furnished to
   others, and derivative works that comment on or otherwise explain it
   or assist in its implementation may be prepared, copied, published
   and distributed, in whole or in part, without restriction of any
   kind, provided that the above copyright notice and this paragraph are
   included on all such copies and derivative works.  However, this
   document itself may not be modified in any way, such as by removing
   the copyright notice or references to the Internet Society or other
   Internet organizations, except as needed for the purpose of
   developing Internet standards in which case the procedures for
   copyrights defined in the Internet Standards process must be
   followed, or as required to translate it into languages other than
   English.

   The limited permissions granted above are perpetual and will not be
   revoked by the Internet Society or its successors or assigns.

   This document and the information contained herein is provided on an
   "AS IS" basis and THE INTERNET SOCIETY AND THE INTERNET ENGINEERING
   TASK FORCE DISCLAIMS 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.

Acknowledgement



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



















Terzis, et al.              Standards Track                    [Page 23]