Network Working Group Tony Przygienda (Xebeo) Internet Draft Naiming Shen (Cisco) <draft-ietf-isis-wg-multi-topology-08.txt> Nischal Sheth (Juniper) March 2005 Expires: September 2005 M-ISIS: Multi Topology (MT) Routing in IS-IS 1. Status of This Memo By submitting this Internet-Draft, we certify that any applicable patent or other IPR claims of which we are aware have been disclosed, or will be disclosed, and any of which we become aware will be disclosed, in accordance with RFC 3668. 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/1id-abstracts.html The list of Internet-Draft Shadow Directories can be accessed at http://www.ietf.org/shadow.html 2. Abstract This draft describes an optional mechanism within ISIS used today by many ISPs for IGP routing within their clouds. This draft describes how to run within a single ISIS domain a set of independent IP topologies that we call Multi-Topologies (MTs). This MT extension can be used for variety of purposes such as an in-band management network ``on top'' of the original IGP topology, maintain separate IGP routing domains for isolated multicast or IPv6 islands within the backbone, or force a subset of an address space to follow a different topology. 3. Specification of Requirements 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]. Przygienda, Shen, Sheth Expires September 2005 [Page 1]
Internet Draft M-ISIS March 2005 4. Introduction Maintaining multiple MTs for ISIS [ISO10589 , RFC1195] in a backwards-compatible manner necessitates several extensions to the packet encoding and additional SPF procedures. The problem can be partitioned into forming of adjacencies, and advertising of prefixes and reachable intermediate systems within each topology. Having put all the necessary additional information in place, it must be properly used by MT capable SPF computation. The following sections describe each of the problems separately. To simplify the text, ``standard'' ISIS topology is defined to be MT ID #0 (zero). 5. Maintaining MT Adjacencies Each adjacency formed MUST be classified as belonging to a set of MTs on the interface. This is achieved by adding a new TLV into IIH packets that advertises which topologies the interface belongs to. If MT #0 is the only MT on the interface, it is optional to advertise it in the new TLV. Thus not including such a TLV in the IIH implies MT ID #0 capability only. Through this exchange of MT capabilities, a router is able to advertise the IS TLVs in LSPs with common MT set over those adjacencies. As a simplifying side-effect, boundaries between levels will be the same for all MTs. In the case of adjacency contains multiple MTs on an interface, and if there exists overlapping IP address space among the topologies, additional mechanism MAY be used to resolve the topology identity of the incoming IP packets on the interface. 5.1. Forming Adjacencies on Point-to-Point Interfaces Adjacencies on point-to-point interfaces are formed as usual with ISIS routers not implementing MT extensions. If local router does not participate in certain MTs, it will not advertise those MTIDs in it's IIHs and thus will not include that neighbor within it's topology based LSPs. On the other hand, if a MTID is not detected in remote side's IIHs, the local router MUST NOT include that neighbor within it's MT LSPs. The local router SHOULD NOT form adjacency if they don't have at least one common MT set over the interface. 5.2. Forming Adjacencies on Broadcast Interfaces On a LAN, all the routers on the LAN which implement the MT extension MAY advertise their MT capability TLV in their IIHs. If there is at least one adjacency on the LAN interface which belongs to this MT, the MT capable router MUST include according MT IS Reachable TLV in its LSP, otherwise it MAY include this MT Przygienda, Shen, Sheth Expires September 2005 [Page 2]
Internet Draft M-ISIS March 2005 IS Reachable TLV in it's LSP if the LAN interface participates in this MT set. Two Routers on a LAN SHALL always establish adjacency regardless whether they have common MT set or not. This is to ensure all the routers on the LAN can correctly elect the same DIS. The IS SHOULD NOT include the MT IS TLV in its LSP if none of the adjacencies on the LAN contains this MT. The DIS, CSNP and PSNP functions are not changed by MT extension. 6. Advertising MT Reachable Intermediate Systems in LSPs A router MUST include within its LSPs in the Reachable Intermediate Systems TLVs only adjacent nodes that are participating in the according topology and advertise such TLVs only if it participates itself in the according topology. Standard Reachable Intermediate Systems TLV is acting here as MT ID #0 equivalent of the newly introduced MT Reachable Intermediate Systems TLV. A router MUST announce the MT IS TLV when there is at least one adjacency on the interface that belongs to this MT, otherwise it MAY announce the MT IS TLV of an adjacency for a given MT if this interface participates in the LAN. Since it is not possible to prevent a router that does not understand MT extensions from being responsible for generation of the according pseudo-node, it is not possible either to introduce special TLVs in the pseudo-node LSPs nor run distinct DIS elections per MT. Therefore, a generated pseudo-node LSP by DIS MUST contain in its IS Reachable TLV all nodes on the LAN as usual regardless of their MT capabilities. In other words, there is no change to the pseudo-node LSP construction. 7. MTs and Overload, Partition and Attached Bits As stated earlier, all MTs share the same set of L1-L2 boundaries and NETs. However, a router could for each of the MTs become potentially partitioned, overloaded and attached independently. To prevent unnecessary complexity, MT extensions does not support partition repair. The overload, partition and attached bits in LSP header only reflect the status of the default topology. Attached bit and overload bit are part of the MT TLV being distributed within a node's LSP fragment zero. Since each adjacency can belong to different MTs, it is possible that some MTs are L2 attached, and others are not on the same router. The overload bit in the MT TLV can be used to signal the topology being overloaded. A MT based system is considered being overloaded if the overload bit in the MT is set. Przygienda, Shen, Sheth Expires September 2005 [Page 3]
Internet Draft M-ISIS March 2005 Route leaking between the levels SHOULD only be performed within the same MT. 8. Advertising MT Specific IP Prefixes Each of the MTs commands its own address space so a new TLV is necessary for prefixes stored in MTs other than MT ID #0. To make the encoding less confusing when same prefixes are present in multiple MTs and accelerate SPF per MT, rather than adding a sub-TLV in TE extensions, a new TLV is introduced for that purpose that closely follows TE encoding [LS01]. 9. MT SPF Computation Each MT MUST run its own instance of the decision process. The pseudo-node LSPs are used by all topologies during computation. Each non-default topology MAY have it's attached bit and overload bit set in the MT TLV. Reverse connectivity check within SPF MUST follow the according MT to assure the bi-directional reachability within the same MT. The results of each computation SHOULD be stored in a separate RIB in normal cases, otherwise overlapping addresses in different topologies could lead to undesirable routing behavior such as forwarding loops. The forwarding logic and configuration need to ensure the same MT is traversed from the source to the destination for packets. The nexthops derived from the MT SPF MUST belong to the adjacencies conforming to the same MT for correct forwarding. It is recommended for the administrators to ensure consistent configuration of all routers in the domain to prevent undesirable forwarding behavior. 10. LSP Flooding An implementation MAY have the option to use additional MT information in the LSP and on the adjacency to reduce some of the unnecessary LSP flooding over the point-to-point links. If a receiving interface and an outgoing interface don't share any common MT set, the implementation MAY have the option not to flood this LSP out on that interface. Since the fragment zero LSP contains the MTID, the MT capability of any LSP can be identified. If the LSP and the adjacencies of an outgoing interface do not share any common MT capability, the implementation MAY have the option not to flood this LSP out on that interface. An implementation MAY want to have the operators to control those optimization base on network topology and environment to ensure the LSP flooding reliability. Przygienda, Shen, Sheth Expires September 2005 [Page 4]
Internet Draft M-ISIS March 2005 When there is an adjacency MT set change over an point-to-point link and the adjacency is in UP state, both sides SHOULD force an exchange of one set of CSNPs over that link. 11. Packet Encoding Three new TLVs are added to support MT extensions. One of them is common for the LSPs and IIHs. Encoding of Intermediate System TLV and IPv4 Reachable Prefixes is tied to traffic engineering extensions [LS01] to simplify the implementation effort. The main reasons we choose using new TLVs instead of using sub-TLVs inside existing TLV type-22 and type-135 are: In many cases, multi-topologies are non-congruent, using sub-TLV approach will not save LSP space; Many sub-TLVs are already being used in TLV type-22, and many more are being proposed while there is a maximum limit on the TLV size, from the existing TLVs; If traffic engineering or some other applications are being applied per topology level later, the new TLVs can automatically inherit the same attributes already defined for the ``standard'' IPv4 topology without going through long standard process to redefine them per topology. 11.1. Multi-Topology TLV TLV number of this TLV is 229. It contains one or more MTs the router is participating in the following structure: x CODE - 229 x LENGTH - total length of the value field, it should be 2 times the number of MT components. x VALUE - one or more 2-byte MT components, structured as follows: No. of Octets +--------------------------------+ |O |A |R |R | MT ID | 2 +--------------------------------+ Bit O represents the OVERLOAD bit for the MT (only valid in LSP fragment zero for MTs other than ID #0, otherwise should be set to 0 on transmission and ignored on receipt.) Bit A represents the ATTACH bit for the MT (only valid in LSP fragment zero for MTs other than ID #0, otherwise should be set to 0 on transmission and ignored on receipt.) Bits R are reserved, should be set to 0 on transmission and ignored on receipt. MT ID is a 12-bit field containing the ID of the topology being announced. Przygienda, Shen, Sheth Expires September 2005 [Page 5]
Internet Draft M-ISIS March 2005 This MT TLV can advertise up to 127 MTs and it can occur multiple times if needed within IIHs and LSP fragment zero. The result MT set should be the union of all the MT TLV occurrence in the packet. Any other ISIS PDU occurrence of this TLV MUST be ignored. Lack of MT TLV in hellos and fragment zero LSP MUST be interpreted as participation of the advertising interface or router in MT ID #0 only. If a router advertises MT TLV, it has to advertise all the MTs it participates in, specifically including topology ID #0 also. 11.2. MT Intermediate Systems TLV TLV number of this TLV is 222. It is aligned with extended IS reachability TLV type 22 beside an additional two bytes in front at the beginning of the TLV. x CODE - 222 x LENGTH - total length of the value field x VALUE - 2-byte MT membership plus the format of extended IS reachability TLV, structured as follows: No. of Octets +--------------------------------+ |R |R |R |R | MT ID | 2 +--------------------------------+ | extended IS TLV format | 11 - 253 +--------------------------------+ . . . . +--------------------------------+ | extended IS TLV format | 11 - 253 +--------------------------------+ Bits R are reserved, should be set to 0 on transmission and ignored on receipt. MT ID is a 12-bit field containing the non-zero MT ID of the topology being announced. The TLV MUST be ignored if the ID is zero. This is to ensure the consistent view of the standard unicast topology. After the 2-byte MT membership format, the MT IS content is in the same format as extended IS TLV, type 22 [LS01]. It can contain up to 23 neighbors of the same MT if no sub-TLVs are used. This TLV can occur multiple times. 11.3. Multi-Topology Reachable IPv4 Prefixes TLV TLV number of this TLV is 235. It is aligned with extended IP reachability TLV type 135 beside an additional two bytes in front. Przygienda, Shen, Sheth Expires September 2005 [Page 6]
Internet Draft M-ISIS March 2005 x CODE - 235 x LENGTH - total length of the value field x VALUE - 2-byte MT membership plus the format of extended extended IP reachability TLV, structured as follows: No. of Octets +--------------------------------+ |R |R |R |R | MT ID | 2 +--------------------------------+ | extended IP TLV format | 5 - 253 +--------------------------------+ . . . . +--------------------------------+ | extended IP TLV format | 5 - 253 +--------------------------------+ Bits R are reserved, should be set to 0 on transmission and ignored on receipt. MT ID is a 12-bit field containing the non-zero ID of the topology being announced. The TLV MUST be ignored if the ID is zero. This is to ensure the consistent view of the standard unicast topology. After the 2-byte MT membership format, the MT IPv4 content is in the same format as extended IP reachability TLV, type 135 [LS01]. This TLV can occur multiple times. 11.4. Multi-Topology Reachable IPv6 Prefixes TLV TLV number of this TLV is 237. It is aligned with IPv6 Reachability TLV type 236 beside an additional two bytes in front. x CODE - 237 x LENGTH - total length of the value field x VALUE - 2-byte MT membership plus the format of IPv6 Reachability TLV, structured as follows: No. of Octets +--------------------------------+ |R |R |R |R | MT ID | 2 +--------------------------------+ | IPv6 Reachability format | 6 - 253 +--------------------------------+ . . +--------------------------------+ | IPv6 Reachability format | 6 - 253 +--------------------------------+ Przygienda, Shen, Sheth Expires September 2005 [Page 7]
Internet Draft M-ISIS March 2005 Bits R are reserved, should be set to 0 on transmission and ignored on receipt. MT ID is a 12-bit field containing the ID of the topology being announced. The TLV MUST be ignored if the ID is zero. After the 2-byte MT membership format, the MT IPv6 context is in the same format as IPv6 Reachability TLV, type 236 [H01]. This TLV can occur multiple times. No attempt is made in this draft to let the M-ISIS IPv6 topology to understand the "normal" IPv6 prefix with TLV type 236. A network SHOULD only run IS-IS either with IPv6 using type 236 or with M-ISIS IPv6 using type 237. 11.5. Reserved MT ID Values Certain MT topologies are assigned to serve pre-determined purposes: - MT ID #0: Equivalent to the ``standard'' topology. - MT ID #1: Reserved for IPv4 in-band management purposes. - MT ID #2: Reserved for IPv6 routing topology. - MT ID #3: Reserved for IPv4 multicast routing topology. - MT ID #4: Reserved for IPv6 multicast routing topology. - MT ID #5-#3995: Reserved for IETF consensus. - MT ID #3996-#4095: Reserved for development, experimental and proprietary features. 12. MT IP Forwarding Considerations Using MT extension for ISIS routing can result in multiple RIBs on the system. The implementation and configuration MUST make sure the IP packets are only traversed through the nodes and links intended for the topologies and applications. In this section we list some of the known considerations for IP forwarding in various MT scenario. Certain deployment scenarios presented here imply different trade-offs in terms of deployment difficulties and advantages obtained. 12.1. Each MT belong to a distinct address family In this case, each MT related routes are installed into a separate RIB. Multiple topologies can share the same ISIS interface on detecting the incoming packet address family. As an example, IPv4 and IPv6 can share the same interface without any further considerations under MT ISIS. Przygienda, Shen, Sheth Expires September 2005 [Page 8]
Internet Draft M-ISIS March 2005 12.2. Some MTs belong to the same address family 12.2.1. Each interface belongs to one and only one MT In this case, MTs can be used to forward packets from the same address family, even with overlapping addresses. Since the MTs have their dedicated interfaces, and those interfaces can be associated with certain MT RIBs and FIBs. 12.2.2. Multiple MTs share an interface with overlapping addresses Some additional mechanism is needed to select the correct RIBs for the incoming IP packets to determine the correct RIB to make a forwarding decision. For example, if the topologies are QoS partitioned, then the DSCP bits in the IP packet header can be utilized to make the decision. Some IP header or even packet data information may be checked to make the forwarding table selection, such as source IP address in the header can be used to determine the desired forwarding behavior. The generic approach of packet to multiple MT RIB mapping over the same inbound interface is outside the scope of this draft. 12.2.3. Multiple MTs share an interface with non-overlapping addresses When there is no overlap in the address space among all the MTs, strictly speaking the destination address space classifies the topology a packet belongs to. It is possible to install routes from different MTs into a shared RIB. As an example of such a deployment, a special ISIS topology can be setup for certain EBGP nexthop addresses. 12.3 Some MTs are not used for forwarding purpose MT in ISIS may be used even if the resulting RIB is not used for forwarding purposes. As an example, multicast RPF check can be performed on a different RIB than the standard unicast RIB albeit an entirely different RIB is used for the multicast forwarding. However, an incoming packet must be still clearly identified as belonging to a unique topology. 13. MT Network Management Considerations When multiple ISIS topologies exist within a domain, some of the routers can be configured to participate in a subset of the MTs in the network. This section discusses some of the options we have to enable operations or the network management stations to access those routers. Przygienda, Shen, Sheth Expires September 2005 [Page 9]
Internet Draft M-ISIS March 2005 13.1. Create dedicated management topology to include all the nodes This approach is to setup a dedicated management topology or 'in-band' management topology. This 'mgmt' topology will include all the routers need to be managed. The computed routes in the topology will be installed into the 'mgmt' RIB. In the condition of the 'mgmt' topology uses a set of non-overlapping address space with the default topology, those 'mgmt' routes can also be optionally installed into the default RIB. The advantages of duplicate 'mgmt' routes in both RIBs include: the network management utilities on the system does not have to be modified to use specific RIB other than the default RIB; the 'mgmt' topology can share the same link with the default topology if so designed. 13.2. Extend the default topology to all the nodes Even in the case default topology is not used on some of the nodes in the IP forwarding, we may want to extend the default topology to those nodes for the purpose of network management. Operators SHOULD set high cost on the links which belong to the extended portion of the default topology. This way the IP data traffic will not be forwarded through those nodes during network topology changes. 14. Acknowledgments The authors would like to thank Andrew Partan, Dino Farinacci, Derek Yeung, Alex Zinin, Stefano Previdi, Heidi Ou, Steve Luong, Mike Shand and Shankar Vemulapalli for the discussion, their review, comments and contributions to this draft. 15. Security Consideration ISIS security applies to the work presented. No specific security issues with the proposed solutions are known. The authentication procedure for ISIS PDUs is the same regardless of MT information inside the ISIS PDUs. 16. References 16.1. Normative References [ISO10589] ISO. Intermediate System to Intermediate System Routing Exchange Protocol for Use in Conjunction with the Protocol for Providing the Connectionless-Mode Network Service. ISO 10589, 1992. Przygienda, Shen, Sheth Expires September 2005 [Page 10]
Internet Draft M-ISIS March 2005 [RFC1195] R. Callon. Use of OSI ISIS for Routing in TCP/IP and Dual Environments. RFC 1195, December 1990. [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate Requirement Levels", BCP 14, RFC 2119, March 1997. 16.2. Informative References [LS01] T. Li and H. Smit. IS-IS Extensions for Traffic Engineering. RFC 3784, March 2005. [H01] C. Hopps. Routing IPv6 with IS-IS. draft-ietf-isis-ipv6-06.txt, May 2004. (work in progress) 17. Authors' Addresses Tony Przygienda Xebeo One Cragwood Road South Plainfield, NJ 07080 Phone: (908) 222 4225 prz@xebeo.com Naiming Shen Cisco Systems 170 W. Tasman Drive San Jose, CA, 95134 USA naiming@cisco.com Nischal Sheth Juniper Networks 1194 North Mathilda Avenue Sunnyvale, CA 94089 USA nsheth@juniper.net 18. Intellectual Property The IETF takes no position regarding the validity or scope of any Intellectual Property Rights or other rights that might be claimed to pertain to the implementation or use of the technology described in this document or the extent to which any license under such rights might or might not be available; nor does it represent that it has made any independent effort to identify any such rights. Information on the procedures with respect to rights in RFC documents can be found in BCP 78 and BCP 79. Copies of IPR disclosures made to the IETF Secretariat and any assurances of licenses to be made available, or the result of an attempt made to obtain a general license or permission for the use Przygienda, Shen, Sheth Expires September 2005 [Page 11]
Internet Draft M-ISIS March 2005 of such proprietary rights by implementers or users of this specification can be obtained from the IETF on-line IPR repository at http://www.ietf.org/ipr. The IETF invites any interested party to bring to its attention any copyrights, patents or patent applications, or other proprietary rights that may cover technology that may be required to implement this standard. Please address the information to the IETF at ietf-ipr@ietf.org. 19. Full Copyright Statement Copyright (C) The Internet Society (2004). This document is subject to the rights, licenses and restrictions contained in BCP 78, and except as set forth therein, the authors retain all their rights. This document and the information contained herein are provided on an "AS IS" basis and THE CONTRIBUTOR, THE ORGANIZATION HE/SHE REPRESENTS OR IS SPONSORED BY (IF ANY), THE INTERNET SOCIETY AND THE INTERNET ENGINEERING TASK FORCE DISCLAIM ALL WARRANTIES, EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO ANY WARRANTY THAT THE USE OF THE INFORMATION HEREIN WILL NOT INFRINGE ANY RIGHTS OR ANY IMPLIED WARRANTIES OF MERCHANTABILITY OR FITNESS FOR A PARTICULAR PURPOSE. Przygienda, Shen, Sheth Expires September 2005 [Page 12]
