Internet-Draft | Area Proxy | March 2023 |
Li, et al. | Expires 13 September 2023 | [Page] |
Link state routing protocols have hierarchical abstraction already built into them. However, when lower levels are used for transit, they must expose their internal topologies to each other, leading to scale issues.¶
To avoid this, this document discusses extensions to the IS-IS routing protocol that would allow level 1 areas to provide transit, yet only inject an abstraction of the level 1 topology into level 2. Each level 1 area is represented as a single level 2 node, thereby enabling greater scale.¶
This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶
Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶
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."¶
This Internet-Draft will expire on 13 September 2023.¶
Copyright (c) 2023 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 (https://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 Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶
The IS-IS routing protocol IS-IS [ISO10589] currently supports a two-level hierarchy of abstraction. The fundamental unit of abstraction is the 'area', which is a (hopefully) connected set of systems running IS-IS at the same level. Level 1, the lowest level, is abstracted by routers that participate in both Level 1 and Level 2, and they inject area information into Level 2. Level 2 systems seeking to access Level 1, use this abstraction to compute the shortest path to the Level 1 area. The full topology database of Level 1 is not injected into Level 2, only a summary of the address space contained within the area, so the scalability of the Level 2 Link State Database (LSDB) is protected.¶
This works well if the Level 1 area is tangential to the Level 2 area. This also works well if there are several routers in both Level 1 and Level 2 and they are adjacent, so Level 2 traffic will never need to transit Level 1 only routers. Level 1 will not contain any Level 2 topology, and Level 2 will only contain area abstractions for Level 1.¶
Unfortunately, this scheme does not work so well if the Level 1 only area needs to provide transit for Level 2 traffic. For Level 2 shortest path first (SPF) computations to work correctly, the transit topology must also appear in the Level 2 LSDB. This implies that all routers that could provide transit, plus any links that might also provide Level 2 transit must also become part of the Level 2 topology. If this is a relatively tiny portion of the Level 1 area, this is not overly painful.¶
However, with today's data center topologies, this is problematic. A common application is to use a Layer 3 Leaf-Spine (L3LS) topology, which is a folded 3-stage Clos [Clos] fabric. It can also be thought of as a complete bipartite graph. In such a topology, the desire is to use Level 1 to contain the routing dynamics of the entire L3LS topology and then to use Level 2 for the remainder of the network. Leaves in the L3LS topology are appropriate for connection outside of the data center itself, so they would provide connectivity for Level 2. If there are multiple connections to Level 2 for redundancy, or other areas, these too would also be made to the leaves in the topology. This creates a difficulty because there are now multiple Level 2 leaves in the topology, with connectivity between the leaves provided by the spines.¶
Following the current rules of IS-IS, all spine routers would necessarily be part of the Level 2 topology, plus all links between a Level 2 leaf and the spines. In the limit, where all leaves need to support Level 2, it implies that the entire L3LS topology becomes part of Level 2. This is seriously problematic as it more than doubles the LSDB held in the L3LS topology and eliminates any benefits of the hierarchy.¶
This document discusses the handling of IP traffic. Supporting MPLS based traffic is a subject for future work.¶
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 BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶
To address this, we propose to completely abstract away the details of the Level 1 area topology within Level 2, making the entire area look like a single proxy system directly connected to all of the area's Level 2 neighbors. By only providing an abstraction of the topology, Level 2's requirement for connectivity can be satisfied without the full overhead of the area's internal topology. It then becomes the responsibility of the Level 1 area to ensure the forwarding connectivity that's advertised.¶
For this discussion, we'll consider a single Level 1 IS-IS area to be the Inside Area, and the remainder of the Level 2 area is the Outside Area. All routers within the Inside Area speak Level 1 and Level 2 IS-IS on all of the links within the topology. We propose to implement Area Proxy by having a Level 2 Proxy Link State Protocol Data Unit (PDU, LSP) that represents the entire Inside Area. We will refer to this as the Proxy LSP. This is the only LSP from the area that will be flooded into the overall Level 2 LSDB.¶
There are four classes of routers that we need to be concerned with in this discussion:¶
All Inside Edge Routers learn the Area Proxy System Identifier from the Area Proxy TLV advertised by the Area Leader and use that as the system identifier in their Level 2 IS-IS Hello PDUs (IIHs) on all Outside interfaces. Outside Edge Routers should then advertise an adjacency to the Area Proxy System Identifier. This allows all Outside Routers to use the Proxy LSP in their SPF computations without seeing the full topology of the Inside Area.¶
Area Proxy functionality assumes that all circuits on Inside Routers are either Level 1-2 circuits within the Inside Area, or Level 2 circuits between Outside Edge Routers and Inside Edge Routers.¶
Area Proxy Boundary multi-access circuits (i.e. Ethernets in LAN mode) with multiple Inside Edge Routers on them are not supported. The Inside Edge Router on any boundary LAN MUST NOT flood Inside Router LSPs on this link. Boundary LANs SHOULD NOT be enabled for Level 1. An Inside Edge Router may be elected the DIS for a Boundary LAN. In this case using the Area Proxy System Id as the basis for the LAN pseudonode identifier could create a collision, so the Insider Edge Router SHOULD compose the pseudonode identifier using its native system identifier. This choice of pseudonode identifier may confuse neighbors with an extremely strict implementation, in which case the Inside Edge Router may be configured with priority 0, causing an Outside Router to be elected DIS.¶
If the Inside Area supports Segment Routing [RFC8402], then all Inside Nodes MUST advertise an SR Global Block (SRGB). The first value of the SRGB advertised by all Inside Nodes MUST start at the same value. The range advertised for the area will be the minimum of all Inside Nodes.¶
To support Segment Routing, the Area Leader will take the global SID information found in the L1 LSDB and convey that to L2 through the Proxy LSP. Prefixes with SID assignments will be copied to the Proxy LSP. Adjacency SIDs for Outside Edge Nodes will be copied to the Proxy LSP.¶
To further extend Segment Routing, it would be helpful to have a segment that refers to the entire Inside Area. This allows a path to refer to an area and have any node within that area accept and forward the packet. In effect, this becomes an anycast SID that is accepted by all Inside Edge Nodes. The information about this SID is distributed in the Area SID Sub-TLV, as part of the Area Leader's Area Proxy TLV (Section 4.3.2). The Inside Edge Nodes MUST establish forwarding based on this SID. The Area Leader SHALL also include the Area SID in the Proxy LSP so that the remainder of L2 can use it for path construction. (Section 4.4.13).¶
All Inside Routers run Level 1-2 IS-IS and must be explicitly instructed to enable the Area Proxy functionality. To signal their readiness to participate in Area Proxy functionality, they will advertise the Area Proxy TLV in their L2 LSP.¶
The Area Proxy TLV serves multiple functions:¶
A node advertises the Area Proxy TLV in fragment 0 of its L2 LSP. Nodes MUST NOT advertise the Area Proxy TLV in a L1 LSP. Nodes MUST ignore the Area Proxy TLV if it is found in a L1 LSP. The Area Proxy TLV is not used in the Proxy LSP. The format of the Area Proxy TLV is:¶
0 1 2 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | TLV Type | TLV Length | Sub-TLVs ... +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+¶
When Outside Routers perform a Level 2 SPF computation, they will use the Proxy LSP for computing a path transiting the Inside Area. Because the topology has been abstracted away, the cost for transiting the Inside Area will be zero.¶
When Inside Routers perform a Level 2 SPF computation, they MUST ignore the Proxy LSP. Further, because these systems do see the Inside Area topology, the link metrics internal to the area are visible. This could lead to different and possibly inconsistent SPF results, potentially leading to forwarding loops.¶
To prevent this, the Inside Routers MUST consider the metrics of links outside of the Inside Area (inter-area metrics) separately from the metrics of the Inside Area links (intra-area metrics). Intra-area metrics MUST be treated as less than any inter-area metric. Thus, if two paths have different total inter-area metrics, the path with the lower inter-area metric would be preferred, regardless of any intra-area metrics involved. However, if two paths have equal inter-area metrics, then the intra-area metrics would be used to compare the paths.¶
Point-to-Point links between two Inside Routers are considered to be Inside Area links. LAN links which have a pseudonode LSP in the Level 1 LSDB are considered to be Inside Area links.¶
The Area Leader will generate a Proxy LSP that will be flooded across the Inside Area. Inside Routers MUST ignore the contents of the Proxy LSP other than for flooding. The Proxy LSP uses the Area Proxy System Identifier as its Source ID.¶
The Area Leader has several responsibilities. First, it MUST inject the Area Proxy System Identifier into the Level 2 LSDB. Second, the Area Leader MUST generate the Proxy LSP for the Inside Area.¶
The Area Leader is selected using the election mechanisms and TLVs described in Dynamic Flooding for IS-IS [I-D.ietf-lsr-dynamic-flooding].¶
If the Area Leader fails, another candidate may become Area Leader and MUST regenerate the Proxy LSP. The failure of the Area Leader is not visible outside of the area and appears to simply be an update of the Proxy LSP.¶
For consistency, all Area Leader candidates SHOULD be configured with the same Proxy System Id, Proxy Hostname, and any other information that may be inserted into the Proxy LSP.¶
The Area Leader is responsible for distributing information about the area to all Inside Nodes. In particular, the Area Leader distributes the Proxy System Id and the Area SID. This is done using two sub-TLVs of the Area Proxy TLV.¶
The Area Proxy System Id Sub-TLV MUST be used by the Area Leader to distribute the Area Proxy System Id. This is an additional system identifier that is used by Inside Nodes and an indication that Area Proxy is active. The format of this sub-TLV is:¶
0 1 2 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | Proxy System ID | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Proxy System Identifier continued | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+¶
The Area Leader MUST advertise the Area Proxy System Identifier Sub-TLV when it observes that all Inside Routers are advertising the Area Proxy TLV. Their advertisements indicate that they are individually ready to perform Area Proxy functionality. The Area Leader then advertises the Area Proxy System Identifier TLV to indicate that the Inside Area MUST enable Area Proxy functionality.¶
Other candidates for Area Leader MAY also advertise the Area Proxy System Identifier when they observe that all Inside Routers are advertising the Area Proxy Router Capability. All candidates advertising the Area Proxy System Identifier TLV MUST be advertising the same system identifier. Multiple proxy system identifiers in a single area is a misconfiguration and each unique occurrence SHOULD be logged.¶
The Area Leader and other candidates for Area Leader MAY withdraw the Area Proxy System Identifier when one or more Inside Routers are not advertising the Area Proxy Router Capability. This will disable Area Proxy functionality. However, before withdrawing the Area Proxy System Identifier, an implementation SHOULD protect against unnecessary churn from transients by delaying the withdrawal. The amount of delay is implementation-dependent.¶
The Area SID Sub-TLV allows the Area Leader to advertise a prefix and SID that represents the entirety of the Inside Area to the Outside Area. This sub-TLV is learned by all of the Inside Edge Nodes who should consume this SID at forwarding time. The Area SID Sub-TLV has the format:¶
0 1 2 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Type | Length | Flags | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | SID/Index/Label (variable) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+ | Prefix Length | Prefix (variable) | +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+¶
where:¶
The Flags octet is defined as follows:¶
0 1 2 3 4 5 6 7 +-+-+-+-+-+-+-+-+ |F|V|L| | +-+-+-+-+-+-+-+-+¶
where:¶
Each Inside Router generates a Level 2 LSP, and the Level 2 LSPs for the Inside Edge Routers will include adjacencies to Outside Edge Routers. Unlike normal Level 2 operations, these LSPs are not advertised outside of the Inside Area and MUST be filtered by all Inside Edge Routers to not be flooded to Outside Routers. Only the Proxy LSP is injected into the overall Level 2 LSDB.¶
The Area Leader uses the Level 2 LSPs generated by the Inside Edge Routers to generate the Proxy LSP. This LSP is originated using the Area Proxy System Identifier. The Area Leader MAY also insert the following additional TLVs into the Proxy LSP for additional information for the Outside Area. LSPs generated by unreachable nodes MUST NOT be considered.¶
The Area Leader SHOULD insert a Protocols Supported TLV (129) [RFC1195] into the Proxy LSP. The values included in the TLV SHOULD be the protocols supported by the Inside Area.¶
The Area Leader SHOULD insert an Area Addresses TLV (1) [ISO10589] into the Proxy LSP.¶
It is RECOMMENDED that the Area Leader insert the Dynamic Hostname TLV (137) [RFC5301] into the Proxy LSP. The contents of the hostname may be specified by configuration. The presence of the hostname helps to simplify debugging the network.¶
The Area Leader MAY insert the IS Neighbors TLV (2) [ISO10589] into the Proxy LSP for Outside Edge Routers. The Area Leader learns of the Outside Edge Routers by examining the LSPs generated by the Inside Edge Routers copying any IS Neighbors TLVs referring to Outside Edge Routers into the Proxy LSP. Since the Outside Edge Routers advertise an adjacency to the Area Proxy System Identifier, this will result in a bi-directional adjacency.¶
An entry for a neighbor in both the IS Neighbors TLV and the Extended IS Neighbors would be functionally redundant, so the Area Leader SHOULD NOT do this.¶
The Area Leader MAY insert the Extended IS Reachability TLV (22) [RFC5305] into the Proxy LSP. The Area Leader SHOULD copy each Extended IS Reachability TLV advertised by an Inside Edge Router about an Outside Edge Router into the Proxy LSP.¶
If the Inside Area supports Segment Routing and Segment Routing selects a SID where the L-Flag is unset, then the Area Lead SHOULD include an Adjacency Segment Identifier sub-TLV (31) [RFC8667] using the selected SID.¶
If the inside area supports SRv6, the Area Leader SHOULD copy the "SRv6 End.X SID" and "SRv6 LAN End.X SID" sub-TLVs of the extended IS reachability TLVs advertised by Inside Edge Routers about Outside Edge Routers.¶
If the inside area supports Traffic Engineering (TE), the Area Leader SHOULD copy TE related sub-TLVs [RFC5305] Section 3 to each Extended IS Reachability TLV in the Proxy LSP.¶
If the Inside Area supports Multi-Topology, then the Area Leader SHOULD copy each Outside Edge Router advertisement that is advertised by an Inside Edge Router in a MT Intermediate Systems TLV into the Proxy LSP.¶
The Area Leader SHOULD insert additional TLVs describing any routing prefixes that should be advertised on behalf of the area. These prefixes may be learned from the Level 1 LSDB, Level 2 LSDB, or redistributed from another routing protocol. This applies to all of various types of TLVs used for prefix advertisement:¶
For TLVs in the Level 1 LSDB, for a given TLV type and prefix, the Area Leader SHOULD select the TLV with the lowest metric and copy that TLV into the Proxy LSP.¶
When examining the Level 2 LSDB for this function, the Area Leader SHOULD only consider TLVs advertised by Inside Routers. Further, for prefixes that represent Boundary links, the Area Leader SHOULD copy all TLVs that have unique sub-TLV contents.¶
If the Inside Area supports Segment Routing and the selected TLV includes a Prefix Segment Identifier sub-TLV (3) [RFC8667], then the sub-TLV SHOULD be copied as well. The P-Flag SHOULD be set in the copy of the sub-TLV to indicate that penultimate hop popping SHOULD NOT be performed for this prefix. The E-Flag SHOULD be reset in the copy of the sub-TLV to indicate that an explicit NULL is not required. The R-Flag SHOULD simply be copied.¶
The Area Leader MAY insert the Router Capability TLV (242) [RFC7981] into the Proxy LSP. If Segment Routing is supported by the inside area, as indicated by the presence of an SRGB being advertised by all Inside Nodes, then the Area Leader SHOULD advertise an SR-Capabilities sub-TLV (2) [RFC8667] with an SRGB. The first value of the SRGB is the same value as the first value advertised by all Inside Nodes. The range advertised for the area will be the minimum of all ranges advertised by Inside Nodes. The Area Leader SHOULD use its own Router Id in the Router Capability TLV.¶
If SRv6 Capability sub-TLV [RFC7981] is advertised by all Inside Routers, the Area Leader should insert an SRv6 Capability sub-TLV in the Router Capability TLV. Each flag in the SRv6 Capability sub-TLV should be set if the flag is set by all Inside Routers.¶
If the Node Maximum SID Depth (MSD) sub-TLV [RFC8491] is advertised by all Inside Routers, the Area Leader should advertise common MSD types and the smallest supported MSD values for each type.¶
If the Inside Area supports multi-topology, then the Area Leader SHOULD insert the Multi-Topology TLV (229) [RFC5120], including the topologies supported by the Inside Nodes.¶
If any Inside Node is advertising the 'O' (Overload) bit for a given topology, then the Area Leader MUST advertise the 'O' bit for that topology. If any Inside Node is advertising the 'A' (Attach) bit for a given topology, then the Area Leader MUST advertise the 'A' bit for that topology.¶
If an Inside Node advertises the SID/Label Binding or Multi-Topology SID/Label Binding SID TLV [RFC8667], then the Area Leader MAY copy the TLV to the Proxy LSP.¶
If the inside area supports SRv6, the Area Leader SHOULD copy all SRv6 locator TLVs [I-D.ietf-lsr-isis-srv6-extensions] advertised by Inside Routers to the Proxy LSP.¶
If the inside area supports TE, the Area Leader SHOULD advertise a TE Router ID TLV (134) [RFC5305] in the Proxy LSP. It SHOULD copy the Shared Risk Link Group (SRLS) TLVs (138) [RFC5307] advertised by Inside Edge Routers about links to Outside Edge Routers.¶
If the inside area supports IPv6 TE, the Area Leader SHOULD advertise an IPv6 TE Router ID TLV (140) [RFC6119] in the Proxy LSP. It SHOULD also copy the IPv6 SRLG TLVs (139) [RFC6119] advertised by Inside Edge Routers about links to Outside Edge Routers.¶
When SR is enabled, it may be useful to advertise an Area SID which will direct traffic to any of the Inside Edge Routers. The information for the Area SID is distributed to all Inside Edge Routers using the Area SID sub-TLV (Section 4.3.2) by the Area Leader.¶
The Area Leader SHOULD advertise the Area SID information in the Proxy LSP as a Node SID as defined in [RFC8667] Section 2.1. The advertisement in the Proxy LSP informs the Outside Area that packets directed to the SID will be forwarded to one of the Inside Edge Nodes and the Area SID will be consumed.¶
Other uses of the Area SID and area SID prefix are outside the scope of this document. Documents which define other use cases for the Area SID MUST specify whether the SID value should be the same or different from that used in support of Area Proxy.¶
The Inside Edge Router has two additional and important functions. First, it MUST generate IIHs that appear to have come from the Area Proxy System Identifier. Second, it MUST filter the L2 LSPs, Partial Sequence Number PDUs (PSNPs), and Complete Sequence Number PDUs (CSNPs) that are being advertised to Outside Routers.¶
The Inside Edge Router has one or more Level 2 interfaces to Outside Routers. These may be identified by explicit configuration or by the fact that they are not also Level 1 circuits. On these Level 2 interfaces, the Inside Edge Router MUST NOT send an IIH until it has learned the Area Proxy System Id from the Area Leader. Then, once it has learned the Area Proxy System Id, it MUST generate its IIHs on the circuit using the Proxy System Id as the source of the IIH.¶
Using the Proxy System Id causes the Outside Router to advertise an adjacency to the Proxy System Id, not to the Inside Edge Router, which supports the proxy function. The normal system id of the Inside Edge Router MUST NOT be used as it will cause unnecessary adjacencies to form and subsequently flap.¶
For the area proxy abstraction to be effective the L2 LSPs generated by the Inside Routers MUST be restricted to the Inside Area. The Inside Routers know which system ids are members of the Inside Area based on the advertisement of the Area Proxy TLV. To prevent unwanted LSP information from escaping the Inside Area, the Inside Edge Router MUST perform filtering of LSP flooding, CSNPs, and PSNPs. Specifically:¶
The authors would like to thank Bruno Decraene and Gunter Van De Velde for their many helpful comments. The authors would also like to thank a small group that wishes to remain anonymous for their valuable contributions.¶
This memo requests that IANA allocate and assign code point 20 from the IS-IS TLV Codepoints registry for the Area Proxy TLV. The registry fields should be: IIH:n, LSP:y, SNP:n, Purge:n.¶
In association with this, this memo requests that IANA create a registry for code points for the sub-TLVs of the Area Proxy TLV.¶
Value | Name | Reference |
---|---|---|
1 | Area Proxy System Identifier | This document |
2 | Area SID | This document |
This document introduces no new security issues. Security of routing within a domain is already addressed as part of the routing protocols themselves. This document proposes no changes to those security architectures.¶