Independent Submission                                           W. Zhou
Request for Comments: 7910                                 Cisco Systems
Category: Informational                                        June 2016
ISSN: 2070-1721


Interoperability between the Virtual Router Redundancy Protocol and PIM

Abstract

   This document introduces VRRP-aware PIM, a redundancy mechanism for
   the Protocol Independent Multicast (PIM) to interoperate with the
   Virtual Router Redundancy Protocol (VRRP).  It allows PIM to track
   VRRP state and to preserve multicast traffic upon failover in a
   redundant network with virtual routing groups enabled.  The mechanism
   described in this document is based on Cisco IOS software
   implementation.

Status of This Memo

   This document is not an Internet Standards Track specification; it is
   published for informational purposes.

   This is a contribution to the RFC Series, independently of any other
   RFC stream.  The RFC Editor has chosen to publish this document at
   its discretion and makes no statement about its value for
   implementation or deployment.  Documents approved for publication by
   the RFC Editor are not a candidate 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
   http://www.rfc-editor.org/info/rfc7910.

Copyright Notice

   Copyright (c) 2016 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
   (http://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.

Table of Contents

   1.  Introduction  . . . . . . . . . . . . . . . . . . . . . . . .   2
   2.  Tracking and Failover . . . . . . . . . . . . . . . . . . . .   3
   3.  PIM Assert Metric Auto-Adjustment . . . . . . . . . . . . . .   4
   4.  DF Election for BiDir Group . . . . . . . . . . . . . . . . .   4
   5.  Tracking Multiple VRRP Groups on an Interface . . . . . . . .   5
   6.  Support of HSRP . . . . . . . . . . . . . . . . . . . . . . .   5
   7.  Security Considerations . . . . . . . . . . . . . . . . . . .   5
   8.  Informative References  . . . . . . . . . . . . . . . . . . .   6
   Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . .   6
   Author's Address  . . . . . . . . . . . . . . . . . . . . . . . .   6

1.  Introduction

   The Virtual Router Redundancy Protocol (VRRP) [RFC5798] is a
   redundancy protocol for establishing a fault-tolerant default router.
   The protocol establishes a framework between network devices in order
   to achieve default device failover if the primary devices become
   inaccessible.

   Protocol Independent Multicast (PIM) [RFC7761] has no inherent
   redundancy capabilities and its operation is completely independent
   of VRRP group states.  As a result, IP multicast traffic is not
   necessarily forwarded by the same device that is elected by VRRP.
   The VRRP-aware PIM feature provides consistent IP multicast
   forwarding in a redundant network with virtual routing groups
   enabled.

   In a multi-access segment (such as LAN), the elected PIM designated
   router (DR) is unaware of the redundancy configuration, and the
   elected DR and VRRP master router (MR) may not be the same.  In order
   to ensure that the PIM DR is always able to forward a PIM Join/Prune
   (J/P) message towards Rendezvous Point (RP) or first-hop router, the
   VRRP MR becomes the PIM DR (if there is only one VRRP group).  PIM is
   responsible for adjusting the DR priority based on the group state.
   When a failover occurs, multicast states are created on the new MR
   elected by the VRRP group and the MR assumes responsibility for the
   routing and forwarding of all the traffic addressed to the VRRP
   virtual IP address (vIP).  This ensures that the PIM DR runs on the
   same router as the VRRP MR and maintains multicast routing (mroute)
   states.  It enables multicast traffic to be forwarded through the
   VRRP MR, allowing PIM to leverage VRRP redundancy, avoid potential
   duplicate traffic, and enable failover, depending on the VRRP states
   in the router.

   This mechanism can be safely deployed into a PIM network without
   changing the behavior of other routers.  When only a specific set of
   routers enable this feature, a user can configure PIM interfaces to
   track state-change events of desired VRRP group(s).  When a router
   becomes the VRRP MR, the PIM component applies the user-defined DR
   priority value to the interface in order to make it PIM DR.  Other
   routers will not break the functionality of this feature, as long as
   their configured DR priority does not conflict with the participating
   routers.  When deployed in a PIM transit network, downstream routers
   should configure the static route to use vIP as the next-hop address
   for PIM J/P messages in order to take advantage of this feature.  If
   dynamic routing is used and the next-hop address is selected by
   unicast routing information as described in [RFC5294], then these
   routes cannot leverage the VRRP redundancy and failover mechanism.
   These downstream routers, however, do not have to support this new
   feature and there is no additional configuration or coordination
   required from a manageability point of view.  This mechanism does not
   change any bit on the wire, and it has been implemented on Cisco IOS
   software.

2.  Tracking and Failover

   Without the mechanisms described in this document, a PIM component
   will send PIM J/P messages with the DR's IP address to upstream
   routers.  A GenID (Generation Identifier) in a PIM Hello message is
   randomly selected when the router boots and remains the same as long
   as the router is up.  A PIM neighbor reboot can easily be detected if
   its GenID is different from before; in this case, the PIM J/P and
   RP-Set information can be redistributed to the rebooted neighbor.
   With the VRRP-aware PIM mechanism enabled, the PIM component listens
   to the state-change notifications from VRRP and automatically adjusts
   the priority of the PIM DR based on the VRRP state and ensures the
   VRRP MR (if there is only one VRRP group) becomes the DR of the LAN.
   If there are multiple VRRP groups, the DR is determined by the user-
   configured priority value.

   Upon failover, the PIM component triggers communication between
   upstream and downstream routers in order to create mroute states on
   the new VRRP MR.  The PIM component sends an additional PIM Hello
   message using the VRRP vIP as the source address for each active VRRP
   group when a router becomes the VRRP MR.  The PIM Hello message with
   a new GenID will trigger other routers to respond to the VRRP
   failover event in the same way as the PIM neighbor reboot event as
   described in [RFC5294].  Specifically, when a downstream router
   receives this PIM Hello message, it will add the source IP address
   (in this case the VRRP vIP) into its PIM neighbor list and
   immediately send triggered PIM J/P messages towards vIP.  Upstream
   routers will process PIM J/P messages based on the VRRP group state.

   If the PIM J/P next-hop address matches the VRRP vIP, only the
   current VRRP MR will process the PIM J/P messages.  This allows all
   PIM J/P messages to reach the VRRP group vIP and minimizes changes
   and configurations at the downstream routers.

   Alternatively, the implementation can choose to have all VRRP passive
   routers maintain mroute states and record the GenID of the current
   MR.  When a passive router becomes the MR, it uses the existing
   mroute states and the recorded MR GenID in its PIM Hello message.
   This will avoid resending PIM J/P messages upon failover and will
   eliminate the requirement of an additional PIM Hello with vIP.  There
   is no change in on-the-wire behavior or in the PIM and VRRP message
   format.

3.  PIM Assert Metric Auto-Adjustment

   It is possible that, after the VRRP MR switches from router A to B, A
   would still forward multicast traffic, which will result in duplicate
   traffic.  The PIM Assert mechanism will kick in because PIM Assert
   with redundancy is enabled.

   o  If there is only one VRRP group, passive routers will send an
      arbitrary penalty metric preference (PIM_ASSERT_INFINITY - 1) and
      make MR the Assert winner.

   o  If there are multiples VRRP groups configured on an interface, the
      Assert metric preference will be (PIM_ASSERT_INFINITY - 1) if and
      only if all VRRP groups are in Passive state.

   o  If there is at least one VRRP group in Active state, then the
      original Assert metric preference will be used.  That is, the
      winner will be selected between routers using their real Assert
      metric preference with at least one active VRRP Group, as if no
      VRRP is involved.

4.  DF Election for BiDir Group

   Change to Designated Forwarder (DF) offer/winner metric is handled
   similarly to PIM Assert handling with VRRP.

   o  If there is only one VRRP group, passive routers will send a large
      penalty metric preference in an offer (PIM_BIDIR_INFINITY_PREF- 1)
      and make MR the DF winner.

   o  If there are multiples VRRP groups configured on an interface, the
      offer metric preference will be (PIM_BIDIR_INFINITY_PREF- 1) if
      and only if all VRRP groups are in Passive state.

   o  If there is at least one VRRP group in Active state, then the
      original offer metric preference to RP will be used.  That is, the
      winner will be selected between routers using their real offer
      metric, as if no VRRP is involved.

5.  Tracking Multiple VRRP Groups on an Interface

   A user can configure a PIM component to track more than one VRRP
   groups on an interface.  This allows other applications to exploit
   PIM/VRRP interoperability to achieve various goals (e.g., load
   balancing).  Since each VRRP group that is configured on an interface
   could be in different states at any moment, the DR priority is
   adjusted.  The PIM Assert metric and PIM BiDir DF metric should be
   adjusted if and only if all VRRP groups that are configured on an
   interface are in Passive (non-Active) states to ensure that
   interfaces with all-passive VRRP groups do not win DR, Assert, and DF
   election.  In other words, the DR, Assert, and DF winners will be
   elected among the interfaces with at least one active VRRP group.

6.  Support of HSRP

   Although there are differences between VRRP and the Hot Standby
   Router Protocol (HSRP) [RFC2281] -- including the number of backup
   (standby) routers, virtual IP address, and timer intervals -- the
   proposed scheme can also enable HSRP-aware PIM with similar failover
   and the tracking mechanism described in this document.

7.  Security Considerations

   The proposed tracking mechanism does not discuss adding
   authentication to the protocols and introduces no new negative impact
   or threats on security to PIM in either SSM (Source-Specific
   Multicast) or ASM (Any-Source Multicast) mode.  Note that VRRP
   messages from malicious nodes could cause unexpected behaviors such
   as multiple MRs and PIM DRs, which are associated with VRRP-specific
   security issues.  To mitigate the vulnerability of frequent VRRP and
   PIM DR state change from malicious attack, an implementation can
   choose to enable VRRP preemption such that a higher-priority VRRP
   backup router does not take over for a lower-priority MR; this will
   reduce the state-change notifications to a PIM component and
   subsequent mroute state changes.  Detailed analysis of PIM and VRRP
   security is provided in [RFC5294] and [RFC5798].

8.  Informative References

   [RFC2281]  Li, T., Cole, B., Morton, P., and D. Li, "Cisco Hot
              Standby Router Protocol (HSRP)", RFC 2281,
              DOI 10.17487/RFC2281, March 1998,
              <http://www.rfc-editor.org/info/rfc2281>.

   [RFC5294]  Savola, P. and J. Lingard, "Host Threats to Protocol
              Independent Multicast (PIM)", RFC 5294,
              DOI 10.17487/RFC5294, August 2008,
              <http://www.rfc-editor.org/info/rfc5294>.

   [RFC5798]  Nadas, S., Ed., "Virtual Router Redundancy Protocol (VRRP)
              Version 3 for IPv4 and IPv6", RFC 5798,
              DOI 10.17487/RFC5798, March 2010,
              <http://www.rfc-editor.org/info/rfc5798>.

   [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, <http://www.rfc-editor.org/info/rfc7761>.

Acknowledgments

   I would like to give a special thank you and appreciation to Stig
   Venaas for his ideas and comments in this document.

Author's Address

   Wei Zhou
   Cisco Systems
   Tasman Drive
   San Jose, CA  95134
   United States

   Email: zhouweiisu@gmail.com