Provider Provisioned VPN WG Hamid Ould-Brahim Internet Draft Nortel Networks Expiration Date: Novembre 2003 Eric C. Rosen Cisco Systems Yakov Rekhter Juniper Networks (Editors) May 2003 Using BGP as an Auto-Discovery Mechanism for Provider-provisioned VPNs draft-ietf-l3vpn-bgpvpn-auto-00.txt Status of this Memo This document is an Internet-Draft and is in full conformance with all provisions of Section 10 of RFC2026 [RFC-2026]. 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. Abstract In any Provider Provisioned-Based VPN (PPVPN) scheme, the Provider Edge (PE) devices attached to a common VPN must exchange certain information as a prerequisite to establish VPN-specific connectivity. The purpose of this draft is to define a BGP based Ould-Brahim, et. al [Page 1] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 auto-discovery mechanism for both layer-2 VPN architectures and layer-3 VPNs ([VPN-VR]). This mechanism is based on the approach used by [RFC2547-bis] for distributing VPN routing information within the service provider(s). Each VPN scheme uses the mechanism to automatically discover the information needed by that particular scheme. 1. Introduction In any Provider Provisioned-Based VPN (PPVPN) scheme, the Provider Edge (PE) devices attached to a common VPN must exchange certain information as a prerequisite to establish VPN-specific connectivity. The purpose of this draft is to define a BGP based auto-discovery mechanism for both layer-2 VPN architectures (i.e., [L2VPN-KOMP], [L2VPN-ROSEN]) and layer-3 VPNs ([VPN-VR]). This mechanism is based on the approach used by [RFC2547-bis] for distributing VPN routing information within the service provider(s). Each VPN scheme uses the mechanism to automatically discover the information needed by that particular scheme. In [RFC2547-bis] based layer-3 VPNs, VPN-specific routes are exchanged, along with the information needed to enable a PE to determine which routes belong to which VRFs. In [VPN-VR], virtual router (VR) addresses must be exchanged, along with the information needed to enable the PEs to determine which VRs are in the same VPN ("membership"), and which of those VRs are to have VPN connectivity ("topology"). Once the VRs are reachable through the tunnels, routes ("reachability") are then exchanged by running existing routing protocols per VPN basis. The BGP-4 multiprotocol extensions are used to carry various information about VPNs for both layer-2 and layer-3 VPN architectures. VPN-specific information associated with the NLRI is encoded either as attributes of the NLRI, or as part of the NLRI itself, or both. 2. Provider Provisioned VPNs Reference Model Both the layer-2 and layer-3 vpns architectures are using a network reference model as illustrated in figure 1. PE PE +--------------+ +--------------+ +--------+ | +----------+ | | +----------+ | +--------+ | VPN-A | | | VPN-A | | | | VPN-A | | | VPN-A | | Sites |--| |Database /| | BGP route | | Database/| |-| sites | +--------+ | |Processing| |<----------->| |Processing| | +--------+ | +----------+ | Distribution| +----------+ | | | | | Ould-Brahim, et al. May 2003 [Page 2] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 +--------+ | +----------+ | | +----------+ | +--------+ | VPN-B | | | VPN-B | | -------- | | VPN-B | | | VPN-B | | Sites |--| |Database /| |-(Backbones)-| | Database/| |-| sites | +--------+ | |Processing| | -------- | |Processing| | +--------+ | +----------+ | | +----------+ | | | | | +--------+ | +----------+ | | +----------+ | +--------+ | VPN-C | | | VPN-C | | | | VPN-C | | | VPN-C | | Sites |--| |Database /| | | | Database/| |-| sites | +--------+ | |Processing| | | |Processing| | +--------+ | +----------+ | | +----------+ | +--------------+ +--------------+ Figure 1: Network based VPN Reference Model It is assumed that the PEs can use BGP to distribute information to each other. This may be via direct IBGP peering, via direct EBGP peering, via multihop BGP peering, through intermediaries such as Route Reflectors, through a chain of intermediate BGP connections, etc. It is assumed also that the PE knows what architecture it is supporting. 3. Carrying VPN information in BGP Multi-Protocol Extension Attributes The BGP-4 multiprotocol extensions are used to carry various information about VPNs for both layer-2 and layer-3 VPN architectures. VPN-specific information associated with the NLRI is encoded either as attributes of the NLRI, or as part of the NLRI itself, or both. The addressing information in the NLRI field is ALWAYS within the VPN address space, and therefore MUST be unique within the VPN. The address specified in the BGP next hop attribute, on the other hand, is in the service provider addressing space. In L3VPNs, the NLRI contains an address prefix which is within the VPN address space, and therefore must be unique within the VPN. 3.1 Carrying Layer-3 VPN Information in BGP-MP This is done as follows. The NLRI is a VPN-IP address or a labeled VPN-IP address. In the case of the virtual router, the NLRI address prefix is an address of one of the virtual routers configured on the PE. Thus this mechanism allows the virtual routers to discover each other, to set up adjacencies and tunnels to each other, etc. In the case of [RFC2547-bis], the NLRI prefix represents a route to an arbitrary system or set of systems within the VPN. Ould-Brahim, et al. May 2003 [Page 3] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 3.2 Carrying Layer-2 VPN Information in BGP-MP The NLRI carries VPN layer-2 addressing information called VPN-L2 address. A VPN-L2 address is composed of a quantity beginning with an 8 bytes Route Distinguisher (RD) field and a variable length quantity encoded according to the layer-2 VPN architecture used. Different layer-2 VPN solutions use the same common AFI, but different SAFI. The AFI indicates that the NLRI is carrying a VPN-l2 address, while the SAFI indicates solution-specific semantics and syntax of the VPN-l2 address that goes after the RD. The RD must be chosen so as it ensures that each NLRI is globally unique (i.e., the same NLRI does not appear in two VPNs). BGP Route target extended community is used to constrain route distribution between PEs. The BGP Next hop carries the service provider tunnel endpoint address. This draft doesn't preclude the use of additional extended community for encoding specific l2vpn parameters. 4. Interpretation of VPN Information in Layer-3 VPNs 4.1 Interpretation of VPN Information in the [RFC2547-bis] model For details, see [RFC2547-bis]. 4.2 Interpretation of VPN Information in the [VPN-VR] model 4.2.1 Membership Discovery The VPN-ID format as defined in [RFC-2685] is used to identify a VPN. All virtual routers that are members of a specific VPN share the same VPN-ID. A VPN-ID is carried in the NLRI to make addresses of VRs globally unique. Making these addresses globally unique is necessary if one uses BGP for VRs' autodiscovery. 4.2.1 Encoding of the VPN-ID in the NLRI For the virtual router model, the VPN-ID is carried within the route distinguisher (RD) field. In order to hold the 7-bytes VPN-ID, the first byte of RD type field is used to indicate the existence of the VPN-ID format. A value of 0x80 in the first byte of RD's type field indicates that the RD field is carrying the VPN-ID format. In this case, the type field range 0x8000-0x80ff will be reserved for the virtual router case. 4.2.1.2 VPN-ID Extended Community Ould-Brahim, et al. May 2003 [Page 4] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 A new extended community is used to carry the VPN-ID format. This attribute is transitive across the Autonomous system boundary. The type field of the VPN-ID extended community is of regular type to be assigned by IANA [BGP-COMM]. The remaining 7 bytes hold the VPN-ID value field as per [RFC-2685]. The BGP UPDATE message will carry information for a single VPN. It is the VPN-ID Extended Community, or more precisely route filtering based on the Extended Community that allows one VR to find out about other VRs in the same VPN. 4.2.2 VPN Topology Information A new extended community is used to indicate different VPN topology values. This attribute is transitive across the Autonomous system boundary. The value of the type field for extended type is assigned by IANA. The first two bytes of the value field (of the remaining 6 bytes) are reserved. The actual topology values are carried within the remaining four bytes. The following topology values are defined: Value Topology Type 1 "Hub" 2 "Spoke" 3 "Mesh" Arbitrary values can also be used to allow specific topologies to be constructed. VPN connectivity between two VRs within the same VPN is achieved if and only if at least one of them is a hub (the other is a hub or a spoke), or if both VRs are part of a full mesh VPN topology. 4.2.3 Tunnel Discovery Network-based VPNs must be implemented through some form of tunneling mechanism, where the packet formats and/or the addressing used within the VPN can be unrelated to that used to route the tunneled packets across the backbone. There are numerous tunneling mechanisms that can be used by a network based VPN (e.g., IP/IP [RFC-2003], GRE tunnels [RFC-1701], IPSec [RFC-2401], and MPLS tunnels [RFC-3031]). Each of these tunnels allows for opaque transport of frames as packet payload across the backbone, with forwarding disjoint from the address fields of the encapsulated packets. A provider edge router may terminate multiple type of tunnels and forward packets between these tunnels and other network interfaces in different ways. BGP can be used to carry tunnel endpoint addresses between edge routers. For scalability purposes, this draft recommends the use of tunneling mechanisms with demultiplexing capabilities such as IPSec, Ould-Brahim, et al. May 2003 [Page 5] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 MPLS, and GRE (with respect to using GRE -the key field, it is no different than just MPLS over GRE, however there is no specification on how to exchange the key field, while there is a specification and implementations on how to exchange the label). Note that IP in IP doesn't have demultiplexing capabilities. The BGP next hop will carry the service provider tunnel endpoint address. As an example, if IPSec is used as tunneling mechanism, the IPSec tunnel remote address will be discovered through BGP, and the actual tunnel establishment is achieved through IPSec signaling protocol. When MPLS tunneling is used, the label carried in the NLRI field is associated with an address of a VR, where the address is carried in the NLRI and is encoded as a VPN-IP address. 5. Interpretation of VPN Information in Layer-2 VPNs The interpretation of the VPN information in L2VPNs is to be specified as part of each L2VPN solution standardized by PPVPN working group. 6. Virtual Router and [RFC2547-bis] Interworking Scenarios Two interwoking scenarios are considered when the network is using both virtual routers and [RFC2547-bis]. The first scenario is a CE- PE relationship between a PE (implementing [RFC2547-bis]), and a VR appearing as a CE to the PE. The connection between the VR, and the PE can be either direct connectivity, or through a tunnel (e.g., IPSec). The second scenario is when a PE is implementing both architectures. In this particular case, a single BGP session configured on the service provider network can be used to advertise either [RFC2547- bis] VPN information or the virtual router related VPN information. From the VR and the [RFC2547-bis] point of view there is complete separation from data path and addressing schemes. However the PE's interfaces are shared between both architectures. A PE implementing only [RFC2547-bis] will not import routes from a BGP UPDATE message containing the VPN-ID extended community. On the other hand, a PE implementing the virtual router architecture will not import routes from a BGP UPDATE message containing the route target extended community attribute. The granularity at which the information is either [RFC2547-bis] related or VR-related is per BGP UPDATE message. Different SAFI numbers are used to indicate that the message carried in BGP multiprotocol extension attributes is to be handled by the VR or [RFC2547-bis] architectures. SAFI number of 128 is used for [RFC2547- bis] related format. A value of 129 for the SAFI number is for the Ould-Brahim, et al. May 2003 [Page 6] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 virtual router (where the NLRI are carrying a labeled prefixes), and a SAFI value of 140 is for non labeled addresses. 7. Scalability Considerations In this section, we briefly summarize the main characteristics of our model with respect to scalability. Recall that the Service Provider network consists of (a) PE routers, (b) BGP Route Reflectors, (c) P routers (which are neither PE routers nor Route Reflectors), and, in the case of multi-provider VPNs, and (d) ASBRs. A PE router, unless it is a Route Reflector should not retain VPN-related information unless it has at least one VPN with an Import Target identical to one of the VPN-related information Route Target attributes. Inbound filtering should be used to cause such information to be discarded. If a new Import Target is later added to one of the PE's VPNs (a "VPN Join" operation), it must then acquire the VPN-related information it may previously have discarded. This can be done using the refresh mechanism described in [BGP- RFSH]. The outbound route filtering mechanism of [BGP-ORF] can also be used to advantage to make the filtering more dynamic. Similarly, if a particular Import Target is no longer present in any of a PE's VPNs (as a result of one or more "VPN Prune" operations), the PE may discard all VPN-related information which, as a result, no longer have any of the PE's VPN's Import Targets as one of their Route Target Attributes. Note that VPN Join and Prune operations are non-disruptive, and do not require any BGP connections to be brought down, as long as the refresh mechanism of [BGP-RFSH] is used. As a result of these distribution rules, no one PE ever needs to maintain all routes for all VPNs; this is an important scalability consideration. Route reflectors can be partitioned among VPNs so that each partition carries routes for only a subset of the VPNs supported by the Service Provider. Thus no single route reflector is required to maintain VPN-related information for all VPNs. For inter-provider VPNs, if multi-hop EBGP is used, then the ASBRs need not maintain and distribute VPN-related information at all. P routers do not maintain any VPN-related information. In order to properly forward VPN traffic, the P routers need only maintain Ould-Brahim, et al. May 2003 [Page 7] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 routes to the PE routers and the ASBRs. As a result, no single component within the Service Provider network has to maintain all the VPN-related information for all the VPNs. So the total capacity of the network to support increasing numbers of VPNs is not limited by the capacity of any individual component. An important consideration to remember is that one may have any number of INDEPENDENT BGP systems carrying VPN-related information. This is unlike the case of the Internet, where the Internet BGP system must carry all the Internet routes. Thus one significant (but perhaps subtle) distinction between the use of BGP for the Internet routing and the use of BGP for distributing VPN-related information, as described in this document is that the former is not amenable to partition, while the latter is. 8. Security Considerations This draft does not introduce any new security considerations to either [VPN-VR] or [RFC2547-bis]. 9. References [BGP-COMM] Ramachandra, Tappan, et al., "BGP Extended Communities Attribute", June 2001, work in progress [BGP-MP] Bates, Chandra, Katz, and Rekhter, "Multiprotocol Extensions for BGP4", February 1998, RFC 2283 [RFC-3107] Rekhter Y, Rosen E., "Carrying Label Information in BGP4", January 2000, RFC3107 [L2VPN-ROSEN] Rosen, E., et al., "An Architecture for L2VPNs", draft-ietf-ppvpn-l2vpn-00.txt, July 2001, work in progress. [L2VPN-KOMP] Kompella, K., et al., "Layer-2 VPNs over Tunnels", draft-kompella-ppvpn-l2vpn-01.txt, work in progress, June 2001, work in progress.. [L2VPN-VKOMP-LASS] Kompella, V., Lasserre, M., et al., "Transparent VLAN Services over MPLS", draft-lasserre-vkompella-ppvpn-vpls-00.txt, work in progress, November 2001. [L2VPN-DTLS] Kompella, K., et. al., "Decoupled Transparent LAN Services", draft-kompella-ppvpn-dtls-00.txt, Ould-Brahim, et al. May 2003 [Page 8] Internet-Draft draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 October 2001, work in progress. [L2VPN-HVPLS] Kandekar, S., et. al., "Hierarchical Virtual Private LAN Service", draft-khandekar-ppvpn-hvpls-mpls-00.txt, November 2001, work in progress. [L2VPN-LPE] Ould-Brahim, H., Chen, M., et al., "VPLS/LPE L2VPNs: Virtual Private LAN Services using Logical PE Architecture", draft-ouldbrahim-l2vpn-lpe-01.txt, October 2001, work in progress. [RFC-3031] Rosen, Viswanathan, and Callon, "Multiprotocol Label Switching Architecture", RFC3031 [RFC-3032] Rosen, Rekhter, Tappan, Farinacci, Fedorkow, Li, and Conta, "MPLS Label Stack Encoding", RFC3032 [RFC-1701] Hanks, S., Li, T., Farinacci, D. and P. Traina, "Generic Routing Encapsulation (GRE)", RFC 1701, October 1994. [RFC-2003] Perkins, C., "IP Encapsulation within IP", RFC 2003, October 1996. [RFC-2026] Bradner, S., "The Internet Standards Process -- Revision 3", RFC2026, October 1996. [RFC-2401] Kent S., Atkinson R., "Security Architecture for the Internet Protocol", RFC2401, November 1998. [RFC-2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", RFC 2119, March 1997. [RFC2547-bis] Rosen E., et al, "BGP/MPLS VPNs", work in progress. [RFC-2685] Fox B., et al, "Virtual Private Networks Identifier", RFC 2685, September 1999. [TLS-TISSA] "BGP/MPLS Layer-2 VPN", draft-tsenevir-bgpl2vpn-01.txt, work in progress, July 2001. [VPN-VR] Ould-Brahim H., et al., "Network based IP VPN Architecture using Virtual Routers", work in progress. 10. Acknowledgments to be supplied. 11. Author's Addresses Ould-Brahim, et al. May 2003 [Page 9] draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 Hamid Ould-Brahim Nortel Networks P O Box 3511 Station C Ottawa, ON K1Y 4H7, Canada Email: hbrahim@nortelnetworks.com Phone: +1 613 765 3418 Bryan Gleeson Tahoe Networks 3052 Orchard Drive San Jose, CA 95134 USA Email: bryan@tahoenetworks.com Peter Ashwood-Smith Nortel Networks P.O. Box 3511 Station C, Ottawa, ON K1Y 4H7, Canada Phone: +1 613 763 4534 Email: petera@nortelnetworks.com Eric C. Rosen Cisco Systems, Inc. 250 Apollo drive Chelmsford, MA, 01824 E-mail: erosen@cisco.com Yakov Rekhter Juniper Networks 1194 N. Mathilda Avenue Sunnyvale, CA 94089 Email: yakov@juniper.net Luyuan Fang AT&T 200 Laurel Avenue Middletown, NJ 07748 Email: Luyuanfang@att.com Phone: +1 (732) 420 1920 Jeremy De Clercq Alcatel Francis Wellesplein 1 B-2018 Antwerpen, Belgium Phone: +32 3 240 47 52 Email: jeremy.de_clercq@alcatel.be Ould-Brahim, et al. May 2003 [Page 10] draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 Riad Hartani Caspian Networks 170 Baytech Drive San Jose, CA 95143 Phone: 408 382 5216 Email: riad@caspiannetworks.com Tissa Senevirathne Force10 Networks 1440 McCarthy Blvd, Milpitas, CA 95035. Phone: 408-965-5103 Email: tsenevir@hotmail.com Ould-Brahim, et al. May 2003 [Page 11] draft-ietf-ppvpn-bgpvpn-auto-05.txt May 2003 Full Copyright Statement Copyright (C) The Internet Society (date). All Rights Reserved. This document and translations of it may be copied and furnished to others, and derivative works that comment on or otherwise explain it or assist in its implementation may be prepared, copied, published and distributed, in whole or in part, without restriction of any kind, provided that the above copyright notice and this paragraph are included on all such copies and derivative works. However, this document itself may not be modified in any way, such as by removing the copyright notice or references to the Internet Society or other Internet organizations, except as needed for the purpose of developing Internet standards in which case the procedures for copyrights defined in the Internet Standards process must be followed, or as required to translate it into languages other than English. The limited permissions granted above are perpetual and will not be revoked by the Internet Society or its successors or assigns. Ould-Brahim, et al. May 2003 [Page 12]