Network Working Group A. Bashandy Internet Draft B. Pithawala Intended status: Standards Track K. Patel Expires: September 2012 Cisco Systems March 2, 2012 Scalable BGP FRR Protection against Edge Node Failure draft-bashandy-bgp-edge-node-frr-02.txt Abstract Consider a BGP free core scenario. Suppose the edge BGP speakers PE1, PE2,..., PEn know about a prefix P/m via the external routers CE1, CE2,..., CEm. If the edge router PEi crashes or becomes totally disconnected from the core, it is desirable for a core router "P" carrying traffic to the failed edge router PEi to immediately restore traffic by re-tunneling packets originally tunneled to PEi and destined to the prefix P/m to one of the other edge routers that advertised P/m, say PEj, until BGP re-converges. In doing so, it is highly desirable to keep the core BGP-free while not imposing restrictions on external connectivity. Thus (1) a core router should not be required to learn any BGP prefix, (2) the size of the forwarding and routing tables in the core routers should be independent of the number of BGP prefixes,(3) there should be no special router (or group of routers) that handles restoring traffic or the need for one router to store the forwarding table of another router, (4) re-routing traffic without waiting for re-convergence must not cause loops, and (4) there should be no restrictions on what edge routers advertise what prefixes. For labeled prefixes, (6) the repairing core router must swap the label stack advertised by the failed edge router PEi for the prefix P/m with the label stack advertised for the same prefix by the edge router PEj before re- tunneling the packet to PEj Status of this Memo This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79. This document may contain material from IETF Documents or IETF Contributions published or made publicly available before November 10, 2008. The person(s) controlling the copyright in some of this material may not have granted the IETF Trust the right to allow modifications of such material outside the IETF Standards Process. Without obtaining an adequate license from the person(s) controlling the copyright in such materials, this document may not be modified outside the IETF Standards Process, and derivative works of it may not be created outside the IETF Standards Process, Bashandy Expires September 2, 2012 [Page 1] Internet-Draft BGP FRR For Edge Node Failure March 2012 except to format it for publication as an RFC or to translate it into languages other than English. Internet-Drafts are working documents of the Internet Engineering Task Force (IETF), its areas, and its working groups. Note that other groups may also distribute working documents as Internet- Drafts. Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress." The list of current Internet-Drafts can be accessed at http://www.ietf.org/ietf/1id-abstracts.txt The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html This Internet-Draft will expire on September 5, 2012. Copyright Notice Copyright (c) 2012 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.........................4 1.2. Terminology...............................................5 1.3. Problem definition........................................6 2. Control Plane Operation........................................7 2.1. Step 1: Calculation of the Repair PE......................8 2.2. Step 2: Assigning and Advertising the BGP Next-hop........8 2.3. Step 3: Informing core routers about the repair path......9 Bashandy Expires September 5, 2012 [Page 2] Internet-Draft BGP FRR For Edge Node Failure March 2012 2.4. Step 4: How a repairing P router (a core router) programs its forwarding plane..............................................10 3. Rules for Choosing and Managing The Repair path...............11 3.1. General Rules for Managing the Repair Path...............11 3.2. Rules for the "Push" Flag................................12 3.3. Rules for Choosing the Repair Path for Labeled Prefixes..13 4. Forwarding Plane Operation....................................14 5. Inter-operability with Existing IP FRR Mechanisms.............15 6. Example.......................................................16 7. Security Considerations.......................................18 8. IANA Considerations...........................................18 9. Conclusions...................................................18 10. References...................................................18 10.1. Normative References....................................18 10.2. Informative References..................................18 11. Acknowledgments..............................................19 Appendix A. Changes from Version 01..............................20 1. Introduction In a BGP free core, where traffic is tunneled between edge routers, BGP speakers advertise reachability information about prefixes to other edge routers not to core rourers. For labeled address families, namely AFI/SAFI 1/4, 2/4, 1/128, and 2/128, an edge router assigns local labels to prefixes and associates the local label with each advertised prefix such as L3VPN [6], 6PE [7], and Softwire [5]. Suppose that a given edge router is chosen as the best next-hop for a prefix P/m. An ingress router that receives a packet from an external router and destined to the prefix P/m "tunnels" the packet across the core to that egress router. If the prefix P/m is a labeled prefix, the ingress router pushes the label advertised by the egress router before tunneling the packet to the egress router. Upon receiving the packet from the core, the egress router takes the appropriate forwarding decision based on the content of the packet or the label pushed on the packet. In modern networks, it is not uncommon to have a prefix reachable via multiple edge routers. One example is the best external path [4]. Another more common and widely deployed scenario is L3VPN [6] with multi-homed VPN sites. As an example, consider the L3VPN topology depicted in Figure 1. Bashandy Expires September 5, 2012 [Page 3] Internet-Draft BGP FRR For Edge Node Failure March 2012 PE1 .............+ | +--------+---------------+ | | | VPN 1 Network | | | | VPN prefix | | (10.0.0.0/8) | | | +---+--------------------+ | /------CE1 / / BGP-free core P--------PE0 \ \ \------CE2 | +---+--------------------+ | | | VPN 2 Network | | | | VPN prefix | | (20.0.0.0/8) | | | +--------+---------------+ | PE2 .............+ Figure 1 VPN prefix reachable via multiple PEs As illustrated in Figure 1, the edge router PE0 is the primary NH for both 10.0.0.0/8 and 20.0.0.0/8. At the same time, both 10.0.0.0/8 and 20.0.0.0/8 are reachable through the other edge routers PE1 and PE2, respectively. 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 [1]. In this document, these words will appear with that interpretation only when in ALL CAPS. Lower case uses of these words are not to be interpreted as carrying RFC-2119 significance. Bashandy Expires September 5, 2012 [Page 4] Internet-Draft BGP FRR For Edge Node Failure March 2012 1.2. Terminology This section defines the terms used in this document. For ease of use, we will use terms similar to those used by L3VPN [6] o BGP-Free core: A network where BGP prefixes are only known to the edge routers and traffic is tunneled between edge routers o Protected prefix: It is a prefix P/m (of any AFI) that a BGP speaker has an external path to. The BGP speaker may learn about the prefix from an external peer through BGP, some other protocol, or manual configuration. The protected prefix is advertised to some or all of the internal peers. o Primary egress PE (Primary PE for simplicity): It is an IBGP peer that can reach the protected prefix P/m through an external path and advertised the prefix to the other IBGP peers. The primary egress PE was chosen as the best path by one or more internal peers. In other words, the primary egress PE is an egress PE that will normally be used by some ingress PEs when there is no failure. Referring to Figure 1, PE0 is a primary egress PE. o Primary next-hop: It is an IPv4 or IPv6 host address belonging to the primary egress PE. If the prefix is advertised via BGP, then the primary next-hop is the next-hop attribute in the BGP update message [2][3]. o CE: It is an external router through which an egress PE can reach a prefix P/m. The routers "CE1" and "E2" in Figure 1 are examples of such CE. o Ingress PE: It is a BGP speaker that learns about a prefix through another IBGP peer and chooses that IBGP peer as the next-hop for the prefix. o Repairing P router (Also "Repairing core router" and "repairing router"): A core router that attempts to restore traffic when the primary egress PE is no longer reachable without waiting for IGP or BGP to re-converge. The repairing P router restores the traffic by rerouting the traffic (through a tunnel) towards the pre-calculated repair PE when it detects that the primary egress PE is no longer reachable. Referring to Figure 1, the router "P" is the repairing P router. Bashandy Expires September 5, 2012 [Page 5] Internet-Draft BGP FRR For Edge Node Failure March 2012 o Repair egress PE (Repair PE for simplicity): It is an egress PE other than the primary egress PE that can reach the protected prefix P/m through an external neighbor. The repair PE is pre- calculated via other PEs prior to any failure. Referring to Figure 1, PE1 is the repair PE for 10.0.0.0/8 while PE2 is the repair PE for 20.0.0.0/8. o Underlying Repair label stack: The underlying repair label stack is the label stack that will be pushed or swapped in when the repairing P router re-tunnels traffic to the repair PE after detecting that the primary egress PE is no longer reachable. o Protected egress PE (Protected PE for simplicity): Any primary egress PE protected by a repairing P router. o Protected edge router: Any protected egress PE. o Repair next-hop: It is an IPv4 or IPv6 host address belonging to the repair egress PE. If the protected prefix is advertised via BGP, then the repair next-hop MAY be the next-hop attribute in the BGP update message [2][3]. o Repair path (Also Repair Egress Path): It is the repair next- hop. If an underlying repair label exists, the repair path is the repair next-hop together with the underlying repair label that will be pushed or swapped in when the repairing P router reroutes traffic to the repair PE. o Primary tunnel: It is the tunnel from the ingress PE to the primary egress PE o Repair tunnel: It is the tunnel from the repairing P router to the repair egress PE 1.3. Problem definition The problem that we are trying to solve is as follows o Even though multiple prefixes may share the same egress router, they have different repair edge router. In Figure 1 above, both 10.0.0.0/8 and 20.0.0.0/8 share the same primary next hop PE0, the routing protocol(s) must identify that the node protecting repair node for 10.0.0.0/8 is PE1 while the node protecting repair node for 11.0.0.0/8 is PE2 Bashandy Expires September 5, 2012 [Page 6] Internet-Draft BGP FRR For Edge Node Failure March 2012 o On loosing connection to the edge router, the core router "P" MUST reroute traffic towards the *correct* repair edge router without waiting for IGP or BGP to re-converge and update the routing tables. On the failure of PE0 illustrated in Figure 1, the core router P needs to reroute traffic for 10.0.0.0/8 towards PE1 and traffic for 11.0.0.0/8 towards PE2 o The repairing core router P MUST NOT be forced to learn about the BGP prefixes on any of the edge router. The same applies for all core routers. o There SHOULD NOT be a need for a special router or group of routers to handle rerouting traffic on edge node failure. Also a router SHOULD not be forced to learn or mirror part or all of the routing or forwarding table of other router(s). o The size of the routing table on any core router MUST be independent of the number of BGP prefixes in the network. o Rerouting traffic without waiting for IGP and BGP to re-converge after a failure MUST NOT cause loops. o For labeled prefixes, repair router MUST forward the re-routed traffic correctly to the external neighbor. Thus the core router MUST either swap the label stack advertised by primary egress edge router (PE0 in Figure 1) with the underlying repair label stack advertised by the repair router (PE1 and PE2 in Figure 1) or push a label on top of the label stack advertised by the primary edge router to allow the repair edge router to forward the packet correctly. 2. Control Plane Operation This section specifies the control plane operation needed to solve the problem defined in the Section 1.3. The control plane operation cover both labeled (AFI/SAFI 1/4, 2/4, 1/128, and 2/128) and unlabeled (AFI/SAFI 1/1, 2/1, 1/2, and 2/2) protected prefixes. The control plane operation can be summarized in 4 steps: 1. Calculation of the repair PE 2. Assigning and advertising the next-hop for protected prefixes 3. Informing core routers about repair paths 4. How a repairing P router (a core router) programs its forwarding plane Bashandy Expires September 5, 2012 [Page 7] Internet-Draft BGP FRR For Edge Node Failure March 2012 2.1. Step 1: Calculation of the Repair PE 1. Consider the prefix P/m learnt by an egress edge router PEi via an external neighbor. If the prefix P/m is labeled, then PEi allocates a local label for each prefix it can reach through an external neighbor CE. A PEi MAY also allocate a repair label such as specified in [11]. 2. Each edge router advertises the unlabeled prefix P/m (belonging to AFI/SAFI 1/1, 2/1, 1/2, and 2/2) and possibly a repair label [11] to some or all of its IBGP peers. If P/m is a labeled prefix (belonging to AFI/SAFI 1/4, 2/4, 1/128, and 2/128), the edge router also advertises the local label. 3. As a result, an edge router PEi having an external path to the prefix P/m may learn about the prefix through other IBGP peers. 4. The edge router PEi MAY chooses a repair PE for the external prefix P/m from among the iBGP peers advertising the prefix P/m. Rules for choosing the repair PE are specified in Section 3. 5. The edge router PEi chooses the one of labels advertised by the other edge router PEj for the prefix P/m as the "underlying repair label". The algorithm for choosing the underlying repair label is specified in Section 3. 6. In the end, if the edge router PEi can reach the prefix P/m through and external path and the prefix P/m is advertised by at least one other PE, the edge router PEi will have o a primary path towards the CE through which the prefix is reachable, and o a repair path consisting of a repair PE and possibly an underlying repair label advertised by the chosen repair PE. 2.2. Step 2: Assigning and Advertising the BGP Next-hop 1. The edge router PEi groups all BGP prefixes for which PEi has an external path and a repair path as follows: a. If the edge router received a repair label from the repair PE: Two prefixes belong to the same group Gi if they share the same repair PE and underlying repair label b. If the edge router did NOT receive a repair label from the repair PE: Two prefixes belong to the same group if they share the same repair PE Bashandy Expires September 5, 2012 [Page 8] Internet-Draft BGP FRR For Edge Node Failure March 2012 2. For each group, the PE assigns a distinct local next-hop. Thus if a prefix P/m belongs to group Gi, its primary next-hop is NHi. 3. When advertising the prefix and its primary next-hop to its IBGP peers, the PE router uses NHi as the next-hop attribute of prefixes belonging to the group Gi. 2.3. Step 3: Informing core routers about the repair path 1. In step 2 (Section 2.2) the egress PE assigns a primary next- hop NHi and a repair path (consisting of a repair next-hop and possibly an underlying repair label) for protected prefixes belonging to group Gi. 1. The primary next-hop NHi is advertised into IGP as usual 2. The repair next-hop is a host address local or advertised by the repair PE. The repair next-hop MAY be the BGP next-hop attribute advertised for the prefix P/m by the repair edge router PEj or some other, possibly future attribute, in the BGP advertisement. Denote the repair next-hop for prefixes belonging to group Gi by "rNHi". Denote the underlying repair label for prefixes belonging to group Gi by "rLi". Because rNHi is the next-hop attribute advertised by the repair PE, rNHi will also be known to all core routers via IGP 3. The repairing egress PE MUST advertise the pair (NHi, rNHi) OR the quadruple (NHi, rNHi, rLi, Push) to core routers that are designated as candidate repairing routers. Note that designating a core router as a candidate repairing router may be subject to administrative actions and/or policy. For example, an administrator may limit candidate repairing routers to only core routers that are directly connected to edge routers. The mechanism by which a core router is designated as a candidate repair router is beyond the scope of this document. 4. The pair (NHi, rNHi) or the quadruple (NHi, rNHi, rLi, Push) may be advertised through various means, such as ISIS optional TLV. The structure and method of advertising (NHi, rNHi) and/or (NHi, rNHi, rLi, Push) is beyond the scope of this document. 5. The semantics of the pair (NHi,rNHi) are: If the next-hop NHi becomes unreachable, then traffic tunneled to the next-hop NHi SHOULD be re-tunneled to the next-hop rNHi because rNHi can reach protected prefixes reachable via the next-hop NHi. Bashandy Expires September 5, 2012 [Page 9] Internet-Draft BGP FRR For Edge Node Failure March 2012 6. The semantics of the quadruple (NHi,r NHi, rLi, Push) are: If the next-hop NHi becomes unreachable, then traffic destined to the next-hop NHi should be re-tunneled to the next-hop rNHi and a. If the PUSH flag is cleared, the label pushed by the ingress PE MUST be swapped with the label rLi before re-tunneling to the repair PE, irrespective of the value of the label pushed by the ingress PE. b. If the PUSH flag is set, then the label rLi MUST be pushed on the packet before re-tunneling to the repair PE. 7. Because of the previous steps, candidate repairing core routers become aware of the repair path for protected BGP prefixes reachable via the primary egress edge router. Note that all core routers remain totally unaware of the BGP prefixes. 2.4. Step 4: How a repairing P router (a core router) programs its forwarding plane 1. Through usual IGP mechanisms, the P router has a prefix matching every BGP next-hop. Let the primary next-hop NHi match the IGP route Ri 2. Any next-hop of prefix Ri is on the path towards the protected egress edge router PEi. A next-hop of the prefix Ri is considered the primary path for the prefix Ri 3. Thus the FIB entry for Ri is programmed as follows a. Primary path: All next hop routers on the path towards NHi b. Repair path when the candidate repair router receives the pair (NHi,rNHi) i. Primary next-hop: the next router on the path towards NHi ii. Repair next-hop: the next-router on the path towards rNHi c. Repair path when the candidate repair router receives the quadruple (NHi,rNHi, rLi, Push) i. If the "Push" flag is *cleared* Pop label in the packet right under the tunnel header (irrespective of the value of that label) Bashandy Expires September 5, 2012 [Page 10] Internet-Draft BGP FRR For Edge Node Failure March 2012 endif ii. Push the underlying repair label rLi iii. Re-tunnel the packet towards the repair next-hop rNHi 3. Rules for Choosing and Managing The Repair path This section specifies rules governing how an egress edge router PEi chooses and advertises the repair path. Other than the rules in this section, the method of choosing the repair path is beyond the scope of this document. 3.1. General Rules for Managing the Repair Path This section specifies general rules for choosing the repair path for both labeled and unlabeled prefixes. 1. A repair PE MUST be another edge router PEj that advertises the same prefix to the primary edge router PEi via IBGP peering. 2. A primary PE MAY advertise more than one repair path for the same primary next-hop NHi. In that case, all advertised repair next-hop identify valid repair PEs for the primary next-hop NHi. Thus a core router MAY choose any repair PE as the repair path for the primary next-hop NHi 3. If a repairing "P" router determines that the path taken by the repair tunnel to a repair edge router PEj passes through the protected edge router PEi, then the repairing router "P" SHOULD NOT install this repair path in its forwarding plane. 4. Let the primary next-hop NHi match the IGP route Ri. If the repairing "P" router determines that the repair tunnel to a repair edge router passes through a next-hop of the IGP route Ri, then the repairing router SHOULD NOT install this repair path in its forwarding plane. 5. A primary next-hop identifies an egress PE. Thus a primary next- hop NH MUST NOT be advertised by two different PEs. However a primary next-hop of one PE MAY be the repair next-hop for another PE. Bashandy Expires September 5, 2012 [Page 11] Internet-Draft BGP FRR For Edge Node Failure March 2012 6. At any point in time, for the same primary and repair next-hops NHi and rNHi, only one advertisement is valid. Thus for the same value of NHi and rNHi, an advertisement of the pair (NHi,rNHi) or the quadruple (NHi,rNHi,rLi,Push) for any values of "rLi" and "Push" MUST override or be preceded by the withdrawal of any previously advertised pair (NHi,rNHi) or the quadruple (NHi,rNHi,rLi,Push). If rules (3) and (4) are not applied, then the tunnel to the repair edge router PEj does not provide protection against the failure of the edge node PEi. Instead it provides core protection against the failure of the path through the core leading to the protected edge node PEi. Thus existing core FRR protection mechanisms such as those specified in [8], [9], and [10] can be used instead. Rules (5) and (6) ensures that there is no ambiguity about the primary and repair next-hops 3.2. Rules for the "Push" Flag This section covers basic rules for advertising, setting, and clearing the "Push" flag 1. If the repairing PE advertises a repair label, then the "Push" flag MUST be advertised. I.e., the repairing PE MUST either advertise the pair (NHi, rNHi) or the quadruple (NHi, rNHi, rLi, Push) 2. If the repair PE advertises a repair label, then the repair PE MAY advertise the "Push" flag with the repair label. In that case, if the primary egress PE (PEi) decides to advertise the quadruple (NHi, rNHi, rLi, Push), then the primary egress PE SHOULD set the "Push" flag to the same value that is received from the repair PE. 3. If the protected prefix is unlabeled (i.e. belongs to AFI/SAFI 1/1, 2/1, 1/2, or 2/2) and the repairing router advertises the quadruple (NHi, rNHi, rLi, Push), then the "Push" flag MUST be set. 4. If the protected prefix is labeled (i.e. belongs to AFI/SAFI 1/4, 2/4, 1/128, and 2/128), then a. The repairing router MUST advertise the quadruple (NHi, rNHi, rLi, Push) b. repairing router MAY set the "Push" flag Bashandy Expires September 5, 2012 [Page 12] Internet-Draft BGP FRR For Edge Node Failure March 2012 Rule (1) is required because the repairing core router(s) need to know what to do with the underlying repair label if it exists Rule (2) is required because the repairing label is allocated by the repairing PE. Hence the repairing PE should be able to specify what to do with it. For example, suppose the repair PE has eiBGP paths for the protected prefix and the protected prefix is unlabeled. In that case, to make sure that the repair PE does not loop the repaired packet back to the primary egress PE, the repair PE advertises a repair label for the unlabeled protected prefix with semantics defined in [11] Rule (3) is needed because the protected prefix is unlabeled. Hence the tunneled packet arriving at the repairing core router "P" has no label and thus the label swapping operation cannot be performed. This document defines the minimum set of rules governing the "Push" flag. Additional rules may be set by other documents. 3.3. Rules for Choosing the Repair Path for Labeled Prefixes This section specifies rules in additions to those mentioned in Section 3.1. by which an egress edge router PEi chooses and advertises the repair path for a protected labeled prefix P/m. 1. A primary edge router PEi SHOULD only choose the edge router PEj and the underlying repair label rLi as a repair path for the prefix P/m if the label advertised for the prefix P/m by the repair edge router PEj is allocated on per-VPN or per-CE/per- next-hop basis. The reason for this is as follows. As mentioned in Section 1.3. the core of the network MUST remain BGP-free and the size of the routing table on a core router MUST remain independent of the number of BGP prefixes. BGP prefix grouping in section 2.2. requires two prefixes to belong to two different groups if the labels advertised by the repair PE for the two prefixes are different. Thus if the repair edge router allocates labels on per-prefix basis, the protected edge router PEi will advertise a different primary next-hop for each protected prefix. This is equivalent to having core router "P" knowing about every BGP prefix. In addition, the size of the routing table of the "P" router becomes comparable to the number of BGP prefixes. 2. If the repair edge router PEj advertises a repair label as described in [11] and the protected edge router understands the repair label attribute described in [11], then the protected edge router PEi SHOULD choose the repair label advertised by PEj as the underlying repair label for the prefix P/m. Bashandy Expires September 5, 2012 [Page 13] Internet-Draft BGP FRR For Edge Node Failure March 2012 Using the repair label specified in [11] has few advantages: o A repairing edge router PEj need not change the primary label allocation policy (which may be per-prefix) but can be chosen as repair PE if the repair labels are allocated on per-CE or per-VRF basis. o As mentioned in [11], an edge router does not repair a packet arriving with a repair label. Hence using the repair label when re-tunneling the packet towards PEj guarantees loop freedom in case of PE-CE link failure. o If the repair PE programs an eiBGP multipath for the protected prefix, then choosing the repair label as specified in [11] guarantees that the repaired packet will not be looped back towards the primary egress PE during repair 4. Forwarding Plane Operation This section specifies the forwarding plane operation on the core router "P" when it detects that the protected edge router PEi is no longer reachable. We assume that the core router has pre-programmed its forwarding plane according to Sections 2.4. As soon as the "P" router detects that the primary next-hop for Ri is not reachable it does the following for any arriving packet destined to the protected edge router PEi 1. Decapsulate the tunnel header to expose the tunneled packet 2. If the underlying repair label rLi is programmed in the forwarding plane a. If the "Push" flag is set Push the underlying repair label rLi b. Else Swap the label on the top of the packet (irrespective of the value of that label) with the underlying repair label rLi 3. Tunnel the packet towards the repair egress PE identified by rNHj Bashandy Expires September 5, 2012 [Page 14] Internet-Draft BGP FRR For Edge Node Failure March 2012 5. Inter-operability with Existing IP FRR Mechanisms Current existing IP FRR mechanisms can be divided into two categories: core protection and edge protection. Core protection techniques, such as [8], [9], and [10], provide protection against internal node and/or link failure. Thus the technique proposed in this document is not related to existing IP FRR mechanisms. If the failure of an internal node or link results in completely disconnecting a protectable edge node, then an administrator MAY configure the repairing router to prefer the technique proposed in this document over existing IP FRR mechanisms. Edge protection techniques, such as [12] and its practical implementation [11] provide protection against the failure of the link between PE and CE routers. Thus existing PE-CE link protection can co-exist with the techniques proposed in this document because the two techniques are independent of each other. Bashandy Expires September 5, 2012 [Page 15] Internet-Draft BGP FRR For Edge Node Failure March 2012 6. Example We will use and LDP core as an example. Consider the diagram depicted in Figure 2 below. We assume that the PEs advertise repair labels as specified in [11] +-----------------------------------+ | | | LDP Core | | | | PE1 | |\ | | \ | | \ | | \ | | CE1....... VPN prefix | | / (10.0.0.0/8) | | / (11.0.0.0/8) | | / | |/ PEx P--------PE0 Lo1 = 1.1.1.1/32 | |\ Lo2 = 2.2.2.2/32 | | \ | | \ | | \ | | CE2....... VPN prefix | | / (20.0.0.0/8) | | / (21.0.0.0/8) | | / | |/ | PE2 | | | | +-----------------------------------+ Figure 2 : Edge node BGP FRR in LDP core 1. As we can see, PE0 has 4 prefixes: 10.0.0.0/8, 11.0.0.0/8, 20.0.0.0/8, and 21.0.0.0/8. PE0 may assign a separate label to each prefix. The method and policy of assigning primary labels to each prefixes is irrelevant to this document. 2. PE1 advertises the repair label rL1 for prefixes 10.0.0.0/8 and 11.0.0.0/8 3. PE2 advertises the repair label rL2 for prefixes 20.0.0.0/8 and 21.0.0.0/8 Bashandy Expires September 5, 2012 [Page 16] Internet-Draft BGP FRR For Edge Node Failure March 2012 4. As such, PE0 divides its prefixes into two groups G1 = {10.0.0.0/8, 11.0.0.0/8} G2 = {20.0.0.0/8, 21.0.0.0/8} 5. When advertising the next-hop to its IBGP peer, PE0 advertises 1.1.1.1 as the next-hop for prefixes belonging to group G1 and 2.2.2.2 as the next-hop for prefixes belonging to group G2. 6. PE0 advertises the prefixes 1.1.1.1/32 and 2.2.2.2/32 using the usual IGP mechanism. 7. When advertising 1.1.1.1/32 into the core, PE0 advertises rL1 and PE1 as a repair path. When advertising 2.2.2.2/32 into the core, PE0 advertises rL2 and PE2 as a repair path. The mechanism by which a repair path is advertised is beyond the scope of the proposal. 8. On the penultimate hop router "P", LDP assigns a different LDP label to 1.1.1.1/32 and 2.2.2.2/32. Core routers other than penultimate hop routers may employ some sort of label aggregation to reduce the number of LDP labels 9. Assume that the penultimate hop router "P" assigns the local LDP label L1 for prefix 1.1.1.1/32 and L2 for prefix 2.2.2.2/32 10.On the penultimate router P, the forwarding entry for L1 will be as follows Primary path: - nexthop is PE0. - swap the incoming outer label with the LDP label towards 1.1.1.1 Repair path - Pop the incoming LDP label - Swap the internal label with the repair label rL1 - Push the LDP label towards PE1 - Forward the packet 11.On the core router P, the forwarding entry for L2 will be as follows Primary path: Same as L1 Repair Path - Pop the incoming LDP label - Swap the internal label with the repair label rL2 - Push the LDP label towards PE2 - Forward the packet Bashandy Expires September 5, 2012 [Page 17] Internet-Draft BGP FRR For Edge Node Failure March 2012 12.If the P router detects that PE0 is no longer reachable, it can use the repair path already pre-programmed in the forwarding plane as described above. Because the repair path is pre- programmed as in the case of TE and IP FRR, the P router can re- route traffic very fast 7. Security Considerations No additional security risk is introduced by using the mechanisms proposed in this document 8. IANA Considerations No requirements for IANA 9. Conclusions This document proposes a method that allows fast re-route protection against edge node failure or complete disconnected from the core in a BGP-free core 10. References 10.1. Normative References [1] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. [2] Rekhter, Y., Li, T., and S. Hares, "A Border Gateway Protocol 4 (BGP-4), RFC 4271, January 2006 [3] Bates, T., Chandra, R., Katz, D., and Rekhter Y., "Multiprotocol Extensions for BGP", RFC 4760, January 2007 10.2. Informative References [4] Marques,P., Fernando, R., Chen, E, Mohapatra, P., Gredler, H., "Advertisement of the best external route in BGP", draft-ietf- idr-best-external-04.txt, April 2011. [5] Wu, J., Cui, Y., Metz, C., and E. Rosen, "Softwire Mesh Framework", RFC 5565, June 2009. [6] Rosen, E. and Y. Rekhter, "BGP/MPLS IP Virtual Private Networks (VPNs)", RFC 4364, February 2006. [7] De Clercq, J. , Ooms, D., Prevost, S., Le Faucheur, F., "Connecting IPv6 Islands over IPv4 MPLS Using IPv6 Provider Edge Routers (6PE)", RFC 4798, February 2007 Bashandy Expires September 5, 2012 [Page 18] Internet-Draft BGP FRR For Edge Node Failure March 2012 [8] Atlas, A. and A. Zinin, "Basic Specification for IP Fast Reroute: Loop-Free Alternates", RFC 5286, September 2008. [9] Shand, S., and Bryant, S., "IP Fast Reroute", RFC5714, January 2010 [10] Shand, M. and S. Bryant, "A Framework for Loop-Free Convergence", RFC 5715, January 2010. [11] Bashandy, A., Pithawala, P., and Heitz, J., "Scalable, Loop- Free BGP FRR using Repair Label", draft-bashandy-idr-bgp- repair-label-02.txt", July 2011 [12] O. Bonaventure, C. Filsfils, and P. Francois. "Achieving sub-50 milliseconds recovery upon bgp peering link failures," IEEE/ACM Transactions on Networking, 15(5):1123-1135, 2007 11. Acknowledgments Special thanks to Les Ginsberg and Anton Smirnov for the valuable comments This document was prepared using 2-Word-v2.0.template.dot. Bashandy Expires September 5, 2012 [Page 19] Internet-Draft BGP FRR For Edge Node Failure March 2012 Appendix A. Changes from Version 01 1. Use the term "underlying repair label" instead of just "repair label" to avoid confusion with the term "repair label" used in [11]. 2. In version 01, it was assumed in many places that the repairing router is the penultimate hop P router. Although this would probably be the most common case, it is not always true. Hence in this version the repairing router may be any core router 3. Merged handling labeled and unlabeled prefixes into a single algorithm. 4. Allowed sending a repair label for unlabeled prefixes and added the "Push" flag. This ensures loop-free repair even for unlabeled prefixes in case that the repair PE has eiBGP paths as mentioned in Section 3.3. 5. In Section 3.3 discussing the rules governing the choice of the underlying repair label for labeled prefix, we changed the wording so that the primary egress PE "SHOULD" instead of "MAY" use the repair label advertised according to [11] as an underlying repair label. 6. All occurrences of the term "backup" were replaced by "repair" as the term "repair" is the commonly used term in the IP FRR context such as [10][9][8] 7. Added the definition of primary and repair tunnels in Section 1.2. 8. Added a definition of the term "Repair Next-hop" in Section 1.2. 9. Modified the definition of "repair path" in Section 1.2 to being the repair next-hop plus the underlying repair label instead of being the repair PE plus the underlying repair label. 10.Outlined inter-operability with existing IP FRR techniques in Section 5. 11.There were few editorial corrections. Bashandy Expires September 5, 2012 [Page 20] Internet-Draft BGP FRR For Edge Node Failure March 2012 Authors' Addresses Ahmed Bashandy Cisco Systems 170 West Tasman Dr, San Jose, CA 95134 Email: bashandy@cisco.com Burjiz Pithawala Cisco Systems 170 West Tasman Dr, San Jose, CA 95134 Email: bpithaw@cisco.com Keyur Patel Cisco Systems 170 West Tasman Dr, San Jose, CA 95134 Email: keyupate@cisco.com Bashandy Expires September 5, 2012 [Page 21]