RFC 8364






Internet Engineering Task Force (IETF)                      IJ. Wijnands
Request for Comments: 8364                                     S. Venaas
Category: Experimental                               Cisco Systems, Inc.
ISSN: 2070-1721                                                  M. Brig
                                                Aegis BMD Program Office
                                                             A. Jonasson
                                                                     FMV
                                                              March 2018


         PIM Flooding Mechanism (PFM) and Source Discovery (SD)

Abstract



   Protocol Independent Multicast - Sparse Mode (PIM-SM) uses a
   Rendezvous Point (RP) and shared trees to forward multicast packets
   from new sources.  Once Last-Hop Routers (LHRs) receive packets from
   a new source, they may join the Shortest Path Tree (SPT) for the
   source for optimal forwarding.  This document defines a new mechanism
   that provides a way to support PIM-SM without the need for PIM
   registers, RPs, or shared trees.  Multicast source information is
   flooded throughout the multicast domain using a new generic PIM
   Flooding Mechanism (PFM).  This allows LHRs to learn about new
   sources without receiving initial data packets.

Status of This Memo



   This document is not an Internet Standards Track specification; it is
   published for examination, experimental implementation, and
   evaluation.

   This document defines an Experimental Protocol for the Internet
   community.  This document is a product of the Internet Engineering
   Task Force (IETF).  It represents the consensus of the IETF
   community.  It has received public review and has been approved for
   publication by the Internet Engineering Steering Group (IESG).  Not
   all documents approved by the IESG are candidates for any level of
   Internet Standard; see Section 2 of RFC 7841.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   https://www.rfc-editor.org/info/rfc8364.









Wijnands, et al.              Experimental                      [Page 1]

RFC 8364                       PFM and SD                     March 2018


Copyright Notice



   Copyright (c) 2018 IETF Trust and the persons identified as the
   document authors.  All rights reserved.

   This document is subject to BCP 78 and the IETF Trust's Legal
   Provisions Relating to IETF Documents
   (https://trustee.ietf.org/license-info) in effect on the date of
   publication of this document.  Please review these documents
   carefully, as they describe your rights and restrictions with respect
   to this document.  Code Components extracted from this document must
   include Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents



   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   3
     1.1.  Conventions Used in This Document . . . . . . . . . . . .   4
     1.2.  Terminology . . . . . . . . . . . . . . . . . . . . . . .   4
   2.  Testing and Deployment Experiences  . . . . . . . . . . . . .   5
   3.  A Generic PIM Flooding Mechanism  . . . . . . . . . . . . . .   5
     3.1.  PFM Message Format  . . . . . . . . . . . . . . . . . . .   6
     3.2.  Administrative Boundaries . . . . . . . . . . . . . . . .   7
     3.3.  Originating PFM Messages  . . . . . . . . . . . . . . . .   7
     3.4.  Processing PFM Messages . . . . . . . . . . . . . . . . .   9
       3.4.1.  Initial Checks  . . . . . . . . . . . . . . . . . . .   9
       3.4.2.  Processing and Forwarding of PFM Messages . . . . . .  10
   4.  Distributing SG Mappings  . . . . . . . . . . . . . . . . . .  11
     4.1.  Group Source Holdtime TLV . . . . . . . . . . . . . . . .  11
     4.2.  Originating Group Source Holdtime TLVs  . . . . . . . . .  12
     4.3.  Processing GSH TLVs . . . . . . . . . . . . . . . . . . .  13
     4.4.  The First Packets and Bursty Sources  . . . . . . . . . .  13
     4.5.  Resiliency to Network Partitioning  . . . . . . . . . . .  14
   5.  Configurable Parameters . . . . . . . . . . . . . . . . . . .  15
   6.  Security Considerations . . . . . . . . . . . . . . . . . . .  15
   7.  IANA Considerations . . . . . . . . . . . . . . . . . . . . .  16
   8.  References  . . . . . . . . . . . . . . . . . . . . . . . . .  16
     8.1.  Normative References  . . . . . . . . . . . . . . . . . .  16
     8.2.  Informative References  . . . . . . . . . . . . . . . . .  17
   Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . .  18
   Authors' Addresses  . . . . . . . . . . . . . . . . . . . . . . .  18









Wijnands, et al.              Experimental                      [Page 2]

RFC 8364                       PFM and SD                     March 2018


1.  Introduction



   Protocol Independent Multicast - Sparse Mode (PIM-SM) [RFC7761] uses
   a Rendezvous Point (RP) and shared trees to forward multicast packets
   to Last-Hop Routers (LHRs).  After the first packet is received by an
   LHR, the source of the multicast stream is learned and the Shortest
   Path Tree (SPT) can be joined.  This document defines a new mechanism
   that provides a way to support PIM-SM without the need for PIM
   registers, RPs, or shared trees.  Multicast source information is
   flooded throughout the multicast domain using a new generic PIM
   flooding mechanism.  By removing the need for RPs and shared trees,
   the PIM-SM procedures are simplified, thus improving router
   operations and management, and making the protocol more robust.
   Also, the data packets are only sent on the SPTs, providing optimal
   forwarding.

   This mechanism has some similarities to Protocol Independent
   Multicast - Dense Mode (PIM-DM) with its State-Refresh signaling
   [RFC3973], except that there is no initial flooding of data packets
   for new sources.  It provides the traffic efficiency of PIM-SM, while
   being as easy to deploy as PIM-DM.  The downside is that it cannot
   provide forwarding of initial packets from a new source, see
   Section 4.4.  PIM-DM is very different from PIM-SM; it's not as
   mature, it is categorized as Experimental not an Internet Standard,
   and there are only a few implementations of it.  The solution in this
   document consists of a lightweight source discovery mechanism on top
   of the Source-Specific Multicast (SSM) [RFC4607] parts of PIM-SM.  It
   is feasible to implement only a subset of PIM-SM to provide SSM
   support and, in addition, implement the mechanism in this document to
   offer a source discovery mechanism for applications that do not
   provide their own source discovery.

   This document defines a generic flooding mechanism for distributing
   information throughout a PIM domain.  While the forwarding rules are
   largely similar to the Bootstrap Router (BSR) mechanism [RFC5059],
   any router can originate information; this allows for flooding of any
   kind of information.  Each message contains one or more pieces of
   information encoded as TLVs.  This document defines one TLV used for
   distributing information about active multicast sources.  Other
   documents may define additional TLVs.

   Note that this document is an Experimental RFC.  While the flooding
   mechanism is largely similar to BSR, there are some concerns about
   scale as there can be multiple routers distributing information, and
   potentially a larger amount of data that needs to be processed and
   stored.  Distributing knowledge of active sources in this way is new;
   there are some concerns, mainly regarding potentially large amounts
   of source states that need to be distributed.  While there has been



Wijnands, et al.              Experimental                      [Page 3]

RFC 8364                       PFM and SD                     March 2018


   some testing in the field, we need to learn more about the forwarding
   efficiency, both the amount of processing per router, propagation
   delay, and the amount of state that can be distributed.  In
   particular, how many active sources one can support without consuming
   too many resources.  There are also parameters, see Section 5, that
   can be tuned regarding how frequently information is distributed.  It
   is not clear what parameters are useful for different types of
   networks.

1.1.  Conventions Used in This Document



   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and
   "OPTIONAL" in this document are to be interpreted as described in
   BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all
   capitals, as shown here.

1.2.  Terminology



   RP:  Rendezvous Point

   BSR:  Bootstrap Router

   RPF:  Reverse Path Forwarding

   SPT:  Shortest Path Tree

   FHR:  First-Hop Router, directly connected to the source

   LHR:  Last-Hop Router, directly connected to the receiver

   PFM:  PIM Flooding Mechanism

   PFM-SD:  PFM Source Discovery

   SG Mapping:  Multicast source group (SG) mapping















Wijnands, et al.              Experimental                      [Page 4]

RFC 8364                       PFM and SD                     March 2018


2.  Testing and Deployment Experiences



   A prototype of this specification has been implemented, and there has
   been some limited testing in the field.  The prototype was tested in
   a network with low-bandwidth radio links.  The network has frequent
   topology changes, including frequent link or router failures.
   Previously existing mechanisms were tested (for example, PIM-SM and
   PIM-DM).

   With PIM-SM, the existing RP election mechanisms were found to be too
   slow.  With PIM-DM, issues were observed with new multicast sources
   starving low-bandwidth links even when there were no receivers; in
   some cases, so much so that there was no bandwidth left for prune
   messages.

   For the PFM-SD prototype tests, all routers were configured to send
   PFM-SD for the directly connected source and to cache received
   announcements.  Applications such as SIP with multicast subscriber
   discovery, multicast voice conferencing, position tracking, and NTP
   were successfully tested.  The tests went quite well.  Packets were
   rerouted as needed; there was no unnecessary forwarding of packets.
   Ease of configuration was seen as a plus.

3.  A Generic PIM Flooding Mechanism



   The Bootstrap Router (BSR) mechanism [RFC5059] is a commonly used
   mechanism for distributing dynamic Group-to-RP mappings in PIM.  It
   is responsible for flooding information about such mappings
   throughout a PIM domain so that all routers in the domain can have
   the same information.  BSR, as defined, is only able to distribute
   Group-to-RP mappings.  This document defines a more generic mechanism
   that can flood any kind of information.  Administrative boundaries,
   see Section 3.2, may be configured to limit to which parts of a
   network the information is flooded.

   The forwarding rules are identical to BSR, except that one can
   control whether routers should forward unsupported data types.  For
   some types of information, it is quite useful that it can be
   distributed without all routers having to support the particular
   type, while there may also be types where it is necessary for every
   single router to support it.  The mechanism includes an originator
   address that is used for RPF checking to restrict the flooding and
   prevent loops, just like BSR.  Like BSR, messages are forwarded hop-
   by-hop; the messages are link-local, and each router will process and
   resend the messages.  Note that there is no equivalent to the BSR
   election mechanism; there can be multiple originators.  This
   mechanism is named the PIM Flooding Mechanism (PFM).




Wijnands, et al.              Experimental                      [Page 5]

RFC 8364                       PFM and SD                     March 2018


3.1.  PFM Message 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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |PIM Ver| Type  |N|  Reserved   |           Checksum            |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |            Originator Address (Encoded-Unicast format)        |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |T|          Type 1             |          Length 1             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                            Value 1                            |
      |                               .                               |
      |                               .                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                               .                               |
      |                               .                               |
      |T|          Type n             |          Length n             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                            Value n                            |
      |                               .                               |
      |                               .                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   PIM Version, Reserved, and Checksum:  As specified in [RFC7761].

   Type:  PIM Message Type.  Value 12 for a PFM message.

   [N]o-Forward bit:  When set, this bit means that the PFM message is
      not to be forwarded.  This bit is defined to prevent Bootstrap
      message forwarding in [RFC5059].

   Originator Address:  The address of the router that originated the
      message.  This can be any address assigned to the originating
      router, but it MUST be routable in the domain to allow successful
      forwarding.  The format for this address is given in the Encoded-
      Unicast address in [RFC7761].

   [T]ransitive bit:  Each TLV in the message includes a bit called the
      "Transitive" bit that controls whether the TLV is forwarded by
      routers that do not support the given type.  See Section 3.4.2.

   Type 1..n:  A message contains one or more TLVs, in this case n TLVs.
      The Type specifies what kind of information is in the Value.  The
      Type range is from 0 to 32767 (15 bits).






Wijnands, et al.              Experimental                      [Page 6]

RFC 8364                       PFM and SD                     March 2018


   Length 1..n:  The length of the Value field in octets.

   Value 1..n:  The value associated with the type and of the specified
      length.

3.2.  Administrative Boundaries



   PFM messages are generally forwarded hop-by-hop to all PIM routers.
   However, similar to BSR, one may configure administrative boundaries
   to limit the information to certain domains or parts of the network.
   Implementations MUST have a way of defining a set of interfaces on a
   router as administrative boundaries for all PFM messages or,
   optionally, for certain TLVs, allowing for different boundaries for
   different TLVs.  Usually, one wants boundaries to be bidirectional,
   but an implementation MAY also provide unidirectional boundaries.
   When forwarding a message, a router MUST NOT send it out on an
   interface that is an outgoing boundary, including a bidirectional
   boundary, for all PFM messages.  If an interface is an outgoing
   boundary for certain TLVs, the message MUST NOT be sent out on the
   interface if it is a boundary for all the TLVs in the message.
   Otherwise, the router MUST remove all the boundary TLVs from the
   message and send the message with the remaining TLVs.  Also, when
   receiving a PFM message on an interface, the message MUST be
   discarded if the interface is an incoming boundary, including a
   bidirectional boundary, for all PFM messages.  If the interface is an
   incoming boundary for certain TLVs, the router MUST ignore all
   boundary TLVs.  If all the TLVs in the message are boundary TLVs,
   then the message is effectively ignored.  Note that when forwarding
   an incoming message, the boundary is applied before forwarding.  If
   the message was discarded or all the TLVs were ignored, then no
   message is forwarded.  When a message is forwarded, it MUST NOT
   contain any TLVs for which the incoming interface is an incoming or
   bidirectional boundary.

3.3.  Originating PFM Messages



   A router originates a PFM message when it needs to distribute
   information using a PFM message to other routers in the network.
   When a message is originated depends on what information is
   distributed.  For instance, this document defines a TLV to distribute
   information about active sources.  When a router has a new active
   source, a PFM message should be sent as soon as possible.  Hence, a
   PFM message should be sent every time there is a new active source.
   However, the TLV also contains a holdtime and PFM messages need to be
   sent periodically.  Generally speaking, a PFM message would typically
   be sent when there is a local state change, causing information to be
   distributed with the PFM to change.  Also, some information may need
   to be sent periodically.  These messages are called "triggered" and



Wijnands, et al.              Experimental                      [Page 7]

RFC 8364                       PFM and SD                     March 2018


   "periodic" messages, respectively.  Each TLV definition will need to
   define when a triggered PFM message needs to be originated, whether
   or not to send periodic messages, and how frequently to send them.

   A router MUST NOT originate more than Max_PFM_Message_Rate messages
   per minute.  This document does not mandate how this should be
   implemented; some possible ways could be having a minimal time
   between each message, counting the number of messages originated and
   resetting the count every minute, or using a leaky bucket algorithm.
   One benefit of using a leaky bucket algorithm is that it can handle
   bursts better.  The default value of Max_PFM_Message_Rate is 6.  The
   value MUST be configurable.  Depending on the network, one may want
   to use a larger value of Max_PFM_Message_Rate to favor propagation of
   new information, but with a large number of routers and many updates,
   the total number of messages might become too large and require too
   much processing.

   There MUST be a minimum of Min_PFM_Message_Gap milliseconds between
   each originated message.  The default value of Min_PFM_Message_Gap is
   1000 (1 second).  The value MUST be configurable.

   Unless otherwise specified by the TLV definitions, there is no
   relationship between different TLVs, and an implementation can choose
   whether to combine TLVs in one message or across separate messages.
   It is RECOMMENDED to combine multiple TLVs in one message to reduce
   the number of messages, but it is also RECOMMENDED that the message
   be small enough to avoid fragmentation at the IP layer.  When a
   triggered PFM message needs to be sent due to a state change, a
   router MAY send a message containing only the information that
   changed.  If there are many changes occurring at about the same time,
   it might be possible to combine multiple changes in one message.  In
   the case where periodic messages are also needed, an implementation
   MAY include periodic PFM information in a triggered PFM.  For
   example, if some information needs to be sent every 60 seconds and a
   triggered PFM message is about to be sent 20 seconds before the next
   periodic PFM message was scheduled, the triggered PFM message might
   include the periodic information and the next periodic PFM message
   can then be scheduled 60 seconds after that rather than 20 seconds
   later.

   When a router originates a PFM message, it puts one of its own
   addresses in the originator field.  An implementation MUST allow an
   administrator to configure which address is used.  For a message to
   be received by all routers in a domain, all the routers need to have
   a route for this address due to the RPF-based forwarding.  Hence, an
   administrator needs to be careful about which address to choose.
   When this is not configured, an implementation MUST NOT use a link-




Wijnands, et al.              Experimental                      [Page 8]

RFC 8364                       PFM and SD                     March 2018


   local address.  It is RECOMMENDED to use an address of a virtual
   interface such that the originator can remain unchanged and routable
   independent of which physical interfaces or links may go down.

   The No-Forward bit MUST NOT be set, except for the case when a router
   receives a PIM Hello from a new neighbor or a PIM Hello with a new
   Generation Identifier (GenID), defined in [RFC7761], is received from
   an existing neighbor.  In that case, an implementation MAY send PFM
   messages containing relevant information so that the neighbor can
   quickly get the correct state.  The definition of the different PFM
   message TLVs needs to specify what, if anything, needs to be sent in
   this case.  If such a PFM message is sent, the No-Forward bit MUST be
   set, and the message must be sent within 60 seconds after the
   neighbor state change.  The processing rules for PFM messages will
   ensure that any other neighbors on the same link ignore the message.
   This behavior (and the choice of 60 seconds) is similar to what is
   defined for the No-Forward bit in [RFC5059].

3.4.  Processing PFM Messages



   A router that receives a PFM message MUST perform the initial checks
   specified here.  If the checks fail, the message MUST be dropped.  An
   error MAY be logged; otherwise, the message MUST be dropped silently.
   If the checks pass, the contents are processed according to the
   processing rules of the included TLVs.

3.4.1.  Initial Checks



   In order to do further processing, a message MUST meet the following
   requirements.  The message MUST be from a directly connected PIM
   neighbor and the destination address MUST be ALL-PIM-ROUTERS.  Also,
   the interface MUST NOT be an incoming, nor a bidirectional,
   administrative boundary for PFM messages, see Section 3.2.  If the
   No-Forward bit is not set, the message MUST be from the RPF neighbor
   of the originator address.  If the No-Forward bit is set, this
   system, the router doing these checks, MUST have enabled the PIM
   protocol within the last 60 seconds.  See Section 3.3 for details.
   In pseudocode, the algorithm is as follows:













Wijnands, et al.              Experimental                      [Page 9]

RFC 8364                       PFM and SD                     March 2018


        if ((DirectlyConnected(PFM.src_ip_address) == FALSE) OR
            (PFM.src_ip_address is not a PIM neighbor) OR
            (PFM.dst_ip_address != ALL-PIM-ROUTERS) OR
            (Incoming interface is admin boundary for PFM)) {
            drop the message silently, optionally log error.
        }
        if (PFM.no_forward_bit == 0) {
            if (PFM.src_ip_address !=
                RPF_neighbor(PFM.originator_ip_address)) {
                drop the message silently, optionally log error.
            }
        } else if (more than 60 seconds elapsed since PIM enabled)) {
            drop the message silently, optionally log error.
        }

   Note that "src_ip_address" is the source address in the IP header of
   the PFM message.  "Originator" is the originator field inside the PFM
   message and is the router that originated the message.  When the
   message is forwarded hop-by-hop, the originator address never
   changes, while the source address will be an address belonging to the
   router that last forwarded the message.

3.4.2.  Processing and Forwarding of PFM Messages



   When the message is received, the initial checks above must be
   performed.  If it passes the checks, then for each included TLV,
   perform processing according to the specification for that TLV.

   After processing, the message is forwarded.  Some TLVs may be omitted
   or modified in the forwarded message.  This depends on administrative
   boundaries (see Section 3.2), the type specification, and the setting
   of the Transitive bit for the TLV.  If a router supports the type,
   then the TLV is forwarded with no changes unless otherwise specified
   by the type specification.  A router not supporting the given type
   MUST include the TLV in the forwarded message if and only if the
   Transitive bit is set.  Whether or not a router supports the type,
   the value of the Transitive bit MUST be preserved if the TLV is
   included in the forwarded message.  The message is forwarded out of
   all interfaces with PIM neighbors (including the interface it was
   received on).  As specified in Section 3.2, if an interface is an
   outgoing boundary for any TLVs, the message MUST NOT be sent out on
   the interface if it is an outgoing boundary for all the TLVs in the
   message.  Otherwise, the router MUST remove any outgoing boundary
   TLVs of the interface from the message and send the message out that
   interface with the remaining TLVs.






Wijnands, et al.              Experimental                     [Page 10]

RFC 8364                       PFM and SD                     March 2018


4.  Distributing SG Mappings



   The generic PFM defined in the previous section can be used for
   distributing SG mappings about active multicast sources throughout a
   PIM domain.  A Group Source Holdtime (GSH) TLV is defined for this
   purpose.

4.1.  Group Source Holdtime TLV



       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
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |1|         Type = 1              |          Length             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |              Group Address (Encoded-Group format)             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |            Src Count          |        Src Holdtime           |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |            Src Address 1 (Encoded-Unicast format)             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |            Src Address 2 (Encoded-Unicast format)             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |                               .                               |
      |                               .                               |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
      |            Src Address m (Encoded-Unicast format)             |
      +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

   1:  The Transitive bit is set to 1.  This means that this type will
      be forwarded even if a router does not support it.  See
      Section 3.4.2.

   Type:  This TLV has type 1.

   Length:  The length of the value in octets.

   Group Address:  The group that sources are to be announced for.  The
      format for this address is given in the Encoded-Group format in
      [RFC7761].

   Src Count:  The number of source addresses that are included.

   Src Holdtime:  The holdtime (in seconds) for the included source(s).

   Src Address:  The source address for the corresponding group.  The
      format for these addresses is given in the Encoded-Unicast address
      in [RFC7761].




Wijnands, et al.              Experimental                     [Page 11]

RFC 8364                       PFM and SD                     March 2018


4.2.  Originating Group Source Holdtime TLVs



   A PFM message MAY contain one or more Group Source Holdtime (GSH)
   TLVs.  This is used to flood information about active multicast
   sources.  Each FHR that is directly connected to an active multicast
   source originates PFM messages containing GSH TLVs.  How a multicast
   router discovers the source of the multicast packet, and when it
   considers itself the FHR, follows the same procedures as the
   registering process described in [RFC7761].  When an FHR has decided
   that a register needs to be sent per [RFC7761], the SG is not
   registered via the PIM-SM register procedures, but the SG mapping is
   included in a GSH TLV in a PFM message.  Note that only the SG
   mapping is distributed in the message: not the entire packet as would
   have been done with a PIM register.

   The PFM messages containing the GSH TLV are sent periodically for as
   long as the multicast source is active, similar to how PIM registers
   are sent periodically.  This means that as long as the source is
   active, it is included in a PFM message originated every
   Group_Source_Holdtime_Period seconds, within the general PFM timing
   requirements in Section 3.3.  The default value of
   Group_Source_Holdtime_Period is 60.  The value MUST be configurable.
   The holdtime for the source MUST be set to either zero or
   Group_Source_Holdtime_Holdtime.  The value of the
   Group_Source_Holdtime_Holdtime parameter MUST be larger than
   Group_Source_Holdtime_Period.  It is RECOMMENDED to be 3.5 times the
   Group_Source_Holdtime_Period.  The default value is 210 (seconds).
   The value MUST be configurable.  A source MAY be announced with a
   holdtime of zero to indicate that the source is no longer active.

   If an implementation supports originating GSH TLVs with different
   holdtimes for different sources, it can (if needed) send multiple
   TLVs with the same group address.  Due to the format, all the sources
   in the same TLV have the same holdtime.

   When a new source is detected, an implementation MAY send a PFM
   message containing just that particular source.  However, it MAY also
   include information about other sources that were just detected,
   sources that are scheduled for periodic announcement later, or other
   types of information.  See Section 3.3 for details.  Note that when a
   new source is detected, one should trigger the sending of a PFM
   message as soon as possible; whereas if a source becomes inactive,
   there is no reason to trigger a message.  There is no urgency in
   removing state for inactive sources.  Note that the message timing
   requirements in Section 3.3 apply.  This means that one cannot always
   send a triggered message immediately when a new source is detected.
   In order to meet the timing requirements, the sending of the message
   may have to be delayed for a small amount of time.



Wijnands, et al.              Experimental                     [Page 12]

RFC 8364                       PFM and SD                     March 2018


   When a new PIM neighbor is detected or an existing neighbor changes
   GenID, an implementation MAY send a triggered PFM message containing
   GSH TLVs for any SG mappings it has learned by receiving PFM GSH TLVs
   as well as any active directly connected sources.  See Section 3.3
   for further details.

4.3.  Processing GSH TLVs



   A router that receives a PFM message containing GSH TLVs MUST parse
   the GSH TLVs and store each of them as SG mappings with an Expiry
   Timer started with the advertised holdtime, that is, unless the
   implementation specifically does not support GSH TLVs, the router is
   configured to ignore GSH TLVs in general, or it is configured to
   ignore GSH TLVs for certain sources or groups.  In particular, an
   administrator might configure a router not to process GSH TLVs if the
   router is known never to have any directly connected receivers.

   For each group that has directly connected receivers, this router
   SHOULD send PIM (S,G) joins for all the SG mappings advertised in the
   message for the group.  Generally, joins are sent, but there could
   be, for instance, an administrative policy limiting which sources and
   groups to join.  The SG mappings are kept alive for as long as the
   Expiry Timer for the source is running.  Once the Expiry Timer
   expires, a PIM router MAY send a PIM (S,G) prune to remove itself
   from the tree.  However, when this happens, there should be no more
   packets sent by the source, so it may be desirable to allow the state
   to time out rather than sending a prune.

   Note that a holdtime of zero has a special meaning.  It is to be
   treated as if the source just expired, and then the state should be
   removed.  Source information MUST NOT be removed due to the source
   being omitted in a message.  For instance, if there are a large
   number of sources for a group, there may be multiple PFM messages,
   each message containing a different list of sources for the group.

4.4.  The First Packets and Bursty Sources



   The PIM register procedure is designed to deliver multicast packets
   to the RP in the absence of an SPT from the RP to the source.  The
   register packets received on the RP are decapsulated and forwarded
   down the shared tree to the LHRs.  As soon as an SPT is built,
   multicast packets would flow natively over the SPT to the RP or LHR
   and the register process would stop.  The PIM register process
   ensures packet delivery until an SPT is in place reaching the FHR.
   If the packets were not unicast encapsulated to the RP, they would be
   dropped by the FHR until the SPT is set up.  This functionality is
   important for applications where the initial packet(s) must be
   received for the application to work correctly.  Another reason would



Wijnands, et al.              Experimental                     [Page 13]

RFC 8364                       PFM and SD                     March 2018


   be for bursty sources.  If the application sends out a multicast
   packet every 4 minutes (or longer), the SPT is torn down (typically
   after 3:30 minutes of inactivity) before the next packet is forwarded
   down the tree.  This will prevent multicast packets from ever being
   forwarded.  A well-behaved application should be able to deal with
   packet loss since IP is a best-effort-based packet delivery system.
   But in reality, this is not always the case.

   With the procedures defined in this document, the packet(s) received
   by the FHR will be dropped until the LHR has learned about the source
   and the SPT is built.  For bursty sources or applications sensitive
   for the delivery of the first packet, that means this solution would
   not be very applicable.  This solution is mostly useful for
   applications that don't have a strong dependency on the initial
   packet(s) and have a fairly constant data rate, like video
   distribution, for example.  For applications with strong dependency
   on the initial packet(s), using BIDIR-PIM [RFC5015] or SSM [RFC4607]
   is recommended.  The protocol operations are much simpler compared to
   PIM-SM; they will cause less churn in the network.  Both guarantee
   best-effort delivery for the initial packet(s).

4.5.  Resiliency to Network Partitioning



   In a PIM-SM deployment where the network becomes partitioned due to
   link or node failure, it is possible that the RP becomes unreachable
   to a certain part of the network.  New sources that become active in
   that partition will not be able to register to the RP and receivers
   within that partition will not be able to receive the traffic.
   Ideally, having a candidate RP in each partition is desirable, but
   which routers will form a partitioned network is something unknown in
   advance.  In order to be fully resilient, each router in the network
   may end up being a candidate RP.  This would increase the operational
   complexity of the network.

   The solution described in this document does not suffer from that
   problem.  If a network becomes partitioned and new sources become
   active, the receivers in that partition will receive the SG mappings
   and join the source tree.  Each partition works independently of the
   other partitions and will continue to have access to sources within
   that partition.  Once the network has healed, the periodic flooding
   of SG mappings ensures that they are reflooded into the other
   partitions and then other receivers can join the newly learned
   sources.








Wijnands, et al.              Experimental                     [Page 14]

RFC 8364                       PFM and SD                     March 2018


5.  Configurable Parameters



   This document contains a number of configurable parameters.  These
   parameters are formally defined in Sections 3.3 and 4.2, but they are
   repeated here for ease of reference.  These parameters all have
   default values as noted below.

   Max_PFM_Message_Rate:  The maximum number of PFM messages a router is
      allowed to originate per minute; see Section 3.3 for details.  The
      default value is 6.

   Min_PFM_Message_Gap:  The minimum amount of time between each PFM
      message originated by a router in milliseconds; see Section 3.3
      for details.  The default is 1000.

   Group_Source_Holdtime_Period:  The announcement period for Group
      Source Holdtime TLVs in seconds; see Section 4.2 for details.  The
      default value is 60.

   Group_Source_Holdtime_Holdtime:  The holdtime for Group Source
      Holdtime TLVs in seconds; see Section 4.2 for details.  The
      default value is 210.

6.  Security Considerations



   For general PIM message security, see [RFC7761].  PFM messages MUST
   only be accepted from a PIM neighbor, but as discussed in [RFC7761],
   any router can become a PIM neighbor by sending a Hello message.  To
   control from where to accept PFM packets, one can limit on which
   interfaces PIM is enabled.  Also, one can configure interfaces as
   administrative boundaries for PFM messages, see Section 3.2.  The
   implications of forged PFM messages depend on which TLVs they
   contain.  Documents defining new TLVs will need to discuss the
   security considerations for the specific TLVs.  In general though,
   the PFM messages are flooded within the network; by forging a large
   number of PFM messages, one might stress all the routers in the
   network.

   If an attacker can forge PFM messages, then such messages may contain
   arbitrary GSH TLVs.  An issue here is that an attacker might send
   such TLVs for a huge amount of sources, potentially causing every
   router in the network to store huge amounts of source state.  Also,
   if there is receiver interest for the groups specified in the GSH
   TLVs, routers with directly connected receivers will build SPTs for
   the announced sources, even if the sources are not actually active.
   Building such trees will consume additional resources on routers that
   the trees pass through.




Wijnands, et al.              Experimental                     [Page 15]

RFC 8364                       PFM and SD                     March 2018


   PIM-SM link-local messages can be authenticated using IPsec, see
   Section 6.3 of [RFC7761] and [RFC5796].  Since PFM messages are link-
   local messages sent hop-by-hop, a link-local PFM message can be
   authenticated using IPsec such that a router can verify that a
   message was sent by a trusted neighbor and has not been modified.
   However, to verify that a received message contains correct
   information announced by the originator specified in the message, one
   will have to trust every router on the path from the originator and
   that each router has authenticated the received message.

7.  IANA Considerations



   This document registers a new PIM message type for the PIM Flooding
   Mechanism (PFM) with the name "PIM Flooding Mechanism" in the "PIM
   Message Types" registry with the value of 12.

   IANA has also created a registry for PFM TLVs called "PIM Flooding
   Mechanism Message Types".  Assignments for the registry are to be
   made according to the policy "IETF Review" as defined in [RFC8126].
   The initial content of the registry is as follows:

      Type         Name                  Reference
      ---------------------------------------------
         0        Reserved               [RFC8364]
         1        Source Group Holdtime  [RFC8364]
      2-32767     Unassigned

8.  References



8.1.  Normative References



   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119,
              DOI 10.17487/RFC2119, March 1997,
              <https://www.rfc-editor.org/info/rfc2119>.

   [RFC5059]  Bhaskar, N., Gall, A., Lingard, J., and S. Venaas,
              "Bootstrap Router (BSR) Mechanism for Protocol Independent
              Multicast (PIM)", RFC 5059, DOI 10.17487/RFC5059, January
              2008, <https://www.rfc-editor.org/info/rfc5059>.

   [RFC5796]  Atwood, W., Islam, S., and M. Siami, "Authentication and
              Confidentiality in Protocol Independent Multicast Sparse
              Mode (PIM-SM) Link-Local Messages", RFC 5796,
              DOI 10.17487/RFC5796, March 2010,
              <https://www.rfc-editor.org/info/rfc5796>.





Wijnands, et al.              Experimental                     [Page 16]

RFC 8364                       PFM and SD                     March 2018


   [RFC7761]  Fenner, B., Handley, M., Holbrook, H., Kouvelas, I.,
              Parekh, R., Zhang, Z., and L. Zheng, "Protocol Independent
              Multicast - Sparse Mode (PIM-SM): Protocol Specification
              (Revised)", STD 83, RFC 7761, DOI 10.17487/RFC7761, March
              2016, <https://www.rfc-editor.org/info/rfc7761>.

   [RFC8126]  Cotton, M., Leiba, B., and T. Narten, "Guidelines for
              Writing an IANA Considerations Section in RFCs", BCP 26,
              RFC 8126, DOI 10.17487/RFC8126, June 2017,
              <https://www.rfc-editor.org/info/rfc8126>.

   [RFC8174]  Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC
              2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174,
              May 2017, <https://www.rfc-editor.org/info/rfc8174>.

8.2.  Informative References



   [RFC3973]  Adams, A., Nicholas, J., and W. Siadak, "Protocol
              Independent Multicast - Dense Mode (PIM-DM): Protocol
              Specification (Revised)", RFC 3973, DOI 10.17487/RFC3973,
              January 2005, <https://www.rfc-editor.org/info/rfc3973>.

   [RFC4607]  Holbrook, H. and B. Cain, "Source-Specific Multicast for
              IP", RFC 4607, DOI 10.17487/RFC4607, August 2006,
              <https://www.rfc-editor.org/info/rfc4607>.

   [RFC5015]  Handley, M., Kouvelas, I., Speakman, T., and L. Vicisano,
              "Bidirectional Protocol Independent Multicast (BIDIR-
              PIM)", RFC 5015, DOI 10.17487/RFC5015, October 2007,
              <https://www.rfc-editor.org/info/rfc5015>.





















Wijnands, et al.              Experimental                     [Page 17]

RFC 8364                       PFM and SD                     March 2018


Acknowledgments

   The authors would like to thank Arjen Boers for contributing to the
   initial idea, and David Black, Stewart Bryant, Yiqun Cai,
   Papadimitriou Dimitri, Toerless Eckert, Dino Farinacci, Alvaro
   Retana, and Liang Xia for their very helpful comments on the
   document.

Authors' Addresses



   IJsbrand Wijnands
   Cisco Systems, Inc.
   De kleetlaan 6a
   Diegem  1831
   Belgium

   Email: ice@cisco.com


   Stig Venaas
   Cisco Systems, Inc.
   Tasman Drive
   San Jose  CA  95134
   United States of America

   Email: stig@cisco.com


   Michael Brig
   Aegis BMD Program Office
   17211 Avenue D, Suite 160
   Dahlgren  VA 22448-5148
   United States of America

   Email: michael.brig@mda.mil


   Anders Jonasson
   Swedish Defence Material Administration (FMV)
   Loennvaegen 4
   Vaexjoe  35243
   Sweden

   Email: anders@jomac.se







Wijnands, et al.              Experimental                     [Page 18]