Request for Comments: 7431
IJ. Wijnands, Ed.
Cisco Systems, Inc.
Multicast-Only Fast Reroute
As IPTV deployments grow in number and size, service providers are looking for solutions that minimize the service disruption due to faults in the IP network carrying the packets for these services. This document describes a mechanism for minimizing packet loss in a network when node or link failures occur. Multicast-only Fast Reroute (MoFRR) works by making simple enhancements to multicast routing protocols such as Protocol Independent Multicast (PIM) and Multipoint LDP (mLDP).
Status of This Memo
This document is not an Internet Standards Track specification; it is published for informational purposes.
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 a candidate for any level of Internet Standard; see Section 2 of RFC 5741.
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/rfc7431.
Copyright © 2015 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. 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 ..........................3 1.2. Terminology ................................................3 2. Basic Overview ..................................................4 3. Determination of the Secondary UMH ..............................5 3.1. ECMP-Mode MoFRR ............................................5 3.2. Non-ECMP-Mode MoFRR ........................................5 4. Upstream Multicast Hop Selection ................................6 4.1. PIM ........................................................6 4.2. mLDP .......................................................6 5. Detecting Failures ..............................................6 6. MoFRR Applicability to Dual-Plane Topology ......................7 7. Other Topologies ...............................................10 8. Capacity Planning for MoFRR ....................................11 9. PE Nodes .......................................................11 10. Other Applications ............................................11 11. Security Considerations .......................................12 12. References ....................................................12 12.1. Normative References .....................................12 12.2. Informative References ...................................12 Acknowledgments ...................................................13 Contributors ......................................................13 Authors' Addresses ................................................14
Different solutions have been developed and deployed to improve service guarantees, both for multicast video traffic and Video on Demand traffic. Most of these solutions are geared towards finding an alternate path around one or more failed network elements (link, node, or path failures).
This document describes a mechanism for minimizing packet loss in a network when node or link failures occur. Multicast-only Fast Reroute (MoFRR) works by making simple changes to the way selected routers use multicast protocols such as PIM and mLDP. No changes to the protocols themselves are required. With MoFRR, in many cases, multicast routing protocols don't necessarily have to depend on or have to wait on unicast routing protocols to detect network failures; see Section 5.
On a Merge Point, MoFRR logic determines a primary Upstream Multicast Hop (UMH) and a secondary UMH and joins the tree via both simultaneously. Data packets are received over the primary and secondary paths. Only the packets from the primary UMH are accepted and forwarded down the tree; the packets from the secondary UMH are discarded. The UMH determination is different for PIM and mLDP and explained in Section 4. When a failure is detected on the path to the primary UMH, the repair occurs by changing the secondary UMH into the primary and the primary into the secondary. Since the repair is local, it is fast -- greatly improving convergence times in the event of node or link failures on the path to the primary UMH.
1.1. Conventions Used in This Document
The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in RFC 2119 [RFC2119].
Multicast-only Fast Reroute.
Multipoint Label Distribution Protocol.
Protocol Independent Multicast.
UMH: Upstream Multicast Hop. A candidate next-hop that can be used
to reach the root of the tree.
tree: Either a PIM (S,G)/(*,G) tree or an mLDP Point-to-Multipoint (P2MP) or Multipoint-to-Multipoint (MP2MP) LSP.
OIF: Outgoing interface. An interface used to forward multicast
packets down the tree towards the receivers. Either a PIM (S,G)/(*,G) tree or an mLDP P2MP or MP2MP LSP.
LFA: Loop-Free Alternate as defined in [RFC5286]. In unicast Fast Reroute, this is an alternate next-hop that can be used to reach a unicast destination without using the protected link or node.
Merge Point: A router that joins a multicast stream via two divergent
Reverse Path Forwarding.
Label Switched Path.
Label Switching Router.
Bidirectional Forwarding Detection.
Interior Gateway Protocol.
Multicast Virtual Private Network.
Point Of Presence, an access point into the network.
2. Basic Overview
The basic idea of MoFRR is for a Merge Point router to join a multicast tree via two divergent upstream paths in order to get maximum redundancy. The determination of this alternate upstream is defined in Section 3.
In order to maximize robustness against any failure, the two paths should be as diverse as possible. Ideally, they should not merge upstream. Sometimes the topology guarantees maximal redundancy; other times additional configuration or techniques are needed to enforce it. See Section 6 for more discussion on the applicability of MoFRR depending on the network topology.
A Merge Point router should only accept and forward on one of the upstream paths at a time in order to avoid duplicate packet forwarding. The selection of the primary and secondary UMH is done by the MoFRR logic and normally based on unicast routing to find loop-free candidates. This is described in Section 4.
Note, the impact of an additional amount of data on the network is mitigated when tree membership is densely populated. When a part of the network has redundant data flowing, join latency for new joining members is reduced because it's likely a tree Merge Point is not far away.
3. Determination of the Secondary UMH
The secondary UMH is a Loop-Free Alternate (LFA) as per [RFC5286].
3.1. ECMP-Mode MoFRR
If the IGP installs two ECMP paths to the source, then as per [RFC5286] the LFA is a primary next-hop. If the multicast tree is enabled for ECMP-mode MoFRR, the router installs the paths as primary and secondary UMHs. Before the failure, only packets received from the primary UMH path are processed, while packets received from the secondary UMH are dropped.
The selected primary UMH SHOULD be the same as if the MoFRR extension were not enabled.
If more than two ECMP paths exist, one is selected as primary and another as secondary UMH. The selection of the primary and secondary is a local decision. Information from the IGP link-state topology could be leveraged to optimize this selection such that the primary and secondary paths are maximal divergent and don't lead to the same upstream node. Note that MoFRR does not restrict the number of UMH paths that are joined. Implementations may use as many paths as are configured.
3.2. Non-ECMP-Mode MoFRR
A router X configured for non-ECMP-mode MoFRR for a multicast tree joins a primary path to its primary UMH and a secondary path to its LFA UMH. In order to prevent control-plane loops, a router MUST stop joining the secondary UMH if this UMH is the only member in the OIF list.
To illustrate the reason for this rule, let's consider the example in Figure 3. If two Provider Edge routers, PE1 and PE2, have received an IGMP request for a multicast tree, they will both join the primary path on their plane and a secondary path to the neighbor PE. If their receivers leave at the same time, it's possible for the multicast tree on PE1 and PE2 to never get deleted, as the PEs refresh each other via the secondary path joins (remember that a secondary path join is not distinguishable from a primary join).
4. Upstream Multicast Hop Selection
An Upstream Multicast Hop (UMH) is a candidate next-hop that can be used to reach the root of the tree. This is normally based on unicast routing to find loop-free candidate(s). With MoFRR procedures, we select a primary and a backup UMH. The procedures for determining the UMH are different for PIM and mLDP.
The UMH selection in PIM is also known as the Reverse Path Forwarding (RPF) procedure. Based on a unicast route lookup on either the source address or Rendezvous Point (RP) [RFC4601], an upstream interface is selected for sending the PIM Joins/Prunes AND accepting the multicast packets. The interface the packets are received on is used to pass or fail the RPF check. If packets are received on an interface that was not selected as the primary by the RPF procedure, the packets are discarded.
The UMH selection in mLDP also depends on unicast routing, but the difference from PIM is that the acceptance of multicast packets is based on MPLS labels and is independent of the interface on which the packet is received. Using the procedures as defined in [RFC6388], an upstream Label Switching Router (LSR) is elected. The upstream LSR that was elected for a Label Switched Path (LSP) gets a unique local MPLS label allocated. Multicast packets are only forwarded if the MPLS label matches the MPLS label that was allocated for that LSP's (primary) upstream LSR.
5. Detecting Failures
Once the two paths are established, the next step is detecting a failure on the primary path to know when to switch to the backup path. This is a local issue, but this section explores some possibilities.
The first (and simplest) option is to detect the failure of the local interface as it's done for unicast Fast Reroute. Detection can be performed using the loss of signal or the loss of probing packets (e.g., BFD). This option can be used in combination with the other options as documented below. Just like for unicast fast reroute, 50 msec switchover is possible.
A second option consists of comparing the packets received on the primary and secondary streams but only forwarding one of them -- the first one received, no matter which interface it is received on. Zero packet loss is possible for RTP-based streams.
A third option assumes a minimum known packet rate for a given data stream. If a packet is not received on the primary RPF within this time frame, the router assumes primary path failure and switches to the secondary RPF interface. 50 msec switchover may be possible for high-rate streams (e.g., IPTV where SD video has a continuous inter- packet gap of about 3 msec), but in general the delay is dependent on the rate of the multicast stream.
A fourth option leverages the significant improvements of the IGP convergence speed. When the primary path to the source is withdrawn by the IGP, the MoFRR-enabled router switches over to the backup path, and the UMH is changed to the secondary UMH. Since the secondary path is already in place, and assuming it is disjoint from the primary path, convergence times would not include the time required to build a new tree and hence are smaller. Sub-second to sub-200 msec switchover should be possible.
6. MoFRR Applicability to Dual-Plane Topology
MoFRR applicability is topology dependent. The applicability is the same as LFA FRR, which is discussed in [RFC6571].
The following section will discuss MoFRR applicability to dual-plane network topologies.
MoFRR works best in dual-planes topologies as illustrated in the figures below. MoFRR may be enabled on any router in the network. In the figures below, MoFRR is shown enabled on the Provider Edge (PE) routers to illustrate one way in which the technology may be deployed.
S P / \ P / \ ^ G1 R1 ^ P / \ P / \ G2----------R2 ^ | \ | \ P ^ | \ | \ P | G3----------R3 | | | | | | | | ^ G4---|------R4 | P ^ \ | \ | P \ | \ | G5----------R5 ^ | | ^ P | | P | | Gi Ri \ \__ ^ /| \ \ S1/ | ^ ^ \ ^\ / |P2 P1 \ S2\_/__ | \ / \| PE1 PE2
P = Primary path
S = Secondary path
Figure 1: Two-Plane Network Design
The topology has two planes, a primary plane and a secondary plane that are fully disjoint from each other all the way into the POPs. This two-plane design is common in service provider networks as it eliminates single point of failures in their core network. The links marked P indicate the normal (primary) path of how the PIM Joins flow from the POPs towards the source of the network. Multicast streams, especially for the densely watched channels, typically flow along both the planes in the network anyway.
The only change MoFRR adds to this is on the links marked S where the PE routers join a secondary path to their secondary ECMP UMH. As a result of this, each PE router receives two copies of the same stream, one from the primary plane and the other from the secondary plane. As a result of normal UMH behavior, the multicast stream received over the primary path is accepted and forwarded to the downstream receivers. The copy of the stream received from the secondary UMH is discarded.
When a router detects a routing failure on the path to its primary UMH, it will switch to the secondary UMH and accept packets for that stream. If the failure is repaired, the router may switch back. The primary and secondary UMHs have only local context and not end-to-end context.
As one can see, MoFRR achieves the faster convergence by pre-building the secondary multicast tree and receiving the traffic on that secondary path. The example discussed above is a simple case where there are two ECMP paths from each PE device towards the source, one along the primary plane and one along the secondary. In cases where the topology is asymmetric or is a ring, this ECMP nature does not hold, and additional rules have to be taken into account to choose when and where to join the secondary path.
MoFRR is appealing in such topologies for the following reasons:
- Ease of deployment and simplicity: the functionality is only required on the PE devices, although it may be configured on all routers in the topology. Furthermore, each PE device can be enabled separately; there is no need for network-wide coordination in order to deploy MoFRR. Interoperability testing is not required as there are no PIM or mLDP protocol changes.
- End-to-end failure detection and recovery: any failure along the path from the source to the PE can be detected and repaired with the secondary disjoint stream. (See the second, third, and fourth options in Section 5.)
- Capacity efficiency: as illustrated in the previous example, the multicast trees corresponding to IPTV channels cover the backbone and distribution topology in a very dense manner. As a consequence, the secondary path grafts onto the normal multicast trees (i.e., trees signaled by PIM or mLDP without the MoFRR extension) at the aggregation level and hence does not demand any extra capacity either on the distribution links or in the backbone. The secondary path simply uses the capacity that is normally used, without any duplication. This is different from conventional FRR mechanisms that often duplicate the capacity requirements when the backup path crosses links/nodes that already carry the primary/normal tree, and thus twice as much capacity is required.
- Loop-free: the secondary path join is sent on an ECMP disjoint path. By definition, the neighbor receiving this request is closer to the source and hence will not cause a loop.
The topology we just analyzed is very frequent and can be modeled as per Figure 2. The PE has two ECMP disjoint paths to the source. Each ECMP path uses a disjoint plane of the network.
Source / \ Plane1 Plane2 | | A1 A2 \ / PE
Figure 2: PE is Dual-Homed to Dual-Plane Backbone
Another frequent topology is described in Figure 3. PEs are grouped by pairs. In each pair, each PE is connected to a different plane. Each PE has one single shortest-path to a source (via its connected plane). There is no ECMP like in Figure 2. However, there is clearly a way to provide MoFRR benefits as each PE can offer a disjoint secondary path to the PE in the other plane (via the disjoint path).
The MoFRR secondary neighbor selection process needs to be extended in this case as one cannot simply rely on using an ECMP path as secondary neighbor. This extension is referred to as non-ECMP-mode MoFRR and is described in Section 3.2.
Source / \ Plane1 Plane2 | | A1 A2 | | PE1----PE2
Figure 3: PEs Are Connected in Pairs to Dual-Plane Backbone
7. Other Topologies
As mentioned in Section 6, MoFRR works best in dual-plane topologies. If MoFRR is applied to non-dual-plane networks, it's possible that the secondary path is affected by the same failure that affected the primary path. In that case, there is no guarantee that the backup path will provide an uninterrupted traffic flow of packets without loss or duplication.
8. Capacity Planning for MoFRR
The previous section has described two very frequent designs (Figures 2 and 3) which provide maximum MoFRR benefits.
Designers with topologies different than Figures 2 and 3 can still benefit from MoFRR, thanks to the use of capacity planning tools.
Such tools are able to simulate the ability of each PE to build two disjoint branches of the same tree. This simulation could be for hundreds of PEs and hundreds of sources.
This allows an assessment of the MoFRR protection coverage of a given network, for a set of sources.
If the protection coverage is deemed insufficient, the designer can use such a tool to optimize the topology (add links, change IGP metrics).
9. PE Nodes
Many Service Providers devise their topology such that PEs have disjoint paths to the multicast sources. MoFRR leverages the existence of these disjoint paths without any PIM or mLDP protocol modification. Interoperability testing is thus not required. In such topologies, MoFRR only needs to be deployed on the PE devices. Each PE device can be enabled one by one.
10. Other Applications
While all the examples in this document show the MoFRR applicability on PE devices, it is clear that MoFRR could be enabled on aggregation or core routers.
MoFRR can be popular in data center network configurations. With the advent of lower-cost Ethernet and increasing port density in routers, there is more meshed connectivity than ever before. When using a three-level access, distribution, and core layers in a data center, there is a lot of inexpensive bandwidth connecting the layers. This will lend itself to more opportunities for ECMP paths at multiple layers. This allows for multiple layers of redundancy protecting link and node failure at each layer with minimal redundancy cost.
Redundancy costs are reduced because only one packet is forwarded at every link along the primary and secondary data paths so there is no duplication of data on any link thereby providing make-before-break protection at a very small cost.
A MoFRR router only accepts packets from the primary path and discards packets from the secondary path. For that reason, management applications (like ping and mtrace) will not work when verifying the secondary path.
The MoFRR principle may be applied to MVPNs.
11. Security Considerations
There are no security considerations for this design other than what is already in the main PIM specification [RFC4601] and mLDP specification [RFC6388].
12.1. 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, <http://www.rfc-editor.org/info/rfc2119>. [RFC5286] Atlas, A., Ed., and A. Zinin, Ed., "Basic Specification for IP Fast Reroute: Loop-Free Alternates", RFC 5286, DOI 10.17487/RFC5286, September 2008, <http://www.rfc-editor.org/info/rfc5286>.
12.2. Informative References
[RFC4601] Fenner, B., Handley, M., Holbrook, H., and I. Kouvelas, "Protocol Independent Multicast - Sparse Mode (PIM-SM): Protocol Specification (Revised)", RFC 4601, DOI 10.17487/RFC4601, August 2006, <http://www.rfc-editor.org/info/rfc4601>. [RFC6388] Wijnands, IJ., Ed., Minei, I., Ed., Kompella, K., and B. Thomas, "Label Distribution Protocol Extensions for Point- to-Multipoint and Multipoint-to-Multipoint Label Switched Paths", RFC 6388, DOI 10.17487/RFC6388, November 2011, <http://www.rfc-editor.org/info/rfc6388>. [RFC6571] Filsfils, C., Ed., Francois, P., Ed., Shand, M., Decraene, B., Uttaro, J., Leymann, N., and M. Horneffer, "Loop-Free Alternate (LFA) Applicability in Service Provider (SP) Networks", RFC 6571, DOI 10.17487/RFC6571, June 2012, <http://www.rfc-editor.org/info/rfc6571>.
Thanks to Dave Oran and Alvaro Retana for their review and comments on this document.
The authors would like to especially acknowledge Dino Farinacci, John Zwiebel, and Greg Shepherd for the genesis of the MoFRR concept.
Below is a list of the contributors in alphabetical order:
Wim Henderickx Alcatel-Lucent Copernicuslaan 50 Antwerp 2018 Belgium Email: email@example.com Uwe Joorde Deutsche Telekom Dahlweg 100 D-48153 Muenster Germany Email: Uwe.Joorde@telekom.de Nicolai Leymann Deutsche Telekom Winterfeldtstrasse 21 Berlin 10781 Germany Email: N.Leymann@telekom.de Jeff Tantsura Ericsson 300 Holger Way San Jose, CA 95134 United States Email: firstname.lastname@example.org
Apoorva Karan Cisco Systems, Inc. 3750 Cisco Way San Jose, CA 95134 United States
email@example.com Clarence Filsfils Cisco Systems, Inc. De kleetlaan 6a Diegem BRABANT 1831 Belgium
firstname.lastname@example.org IJsbrand Wijnands (editor) Cisco Systems, Inc. De Kleetlaan 6a Diegem 1831 Belgium
email@example.com Bruno Decraene Orange 38-40 rue du General Leclerc Issy Moulineaux Cedex 9, 92794 France Email: firstname.lastname@example.org