| Internet-Draft | SRH Insertion | November 2020 |
| Voyer, et al. | Expires 24 May 2021 | [Page] |
SRv6 is deployed in multiple provider networks.¶
This document describes the usage of SRH insertion and deletion within the SR domain and how security and end-to-end integrity is guaranteed.¶
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].¶
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 24 May 2021.¶
Copyright (c) 2020 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 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.¶
[I-D.matsushima-spring-srv6-deployment-status] records multiple SRv6 deployments in multiple networks¶
In each deployment, traffic traversing an SR domain is encapsulated in an outer IPv6 header for its journey through the SR domain.¶
To implement transport services within the SR domain, insertion or removal of an SRH after the outer IPv6 header is performed. Any segment within the SRH is strictly contained within the SR domain.¶
The SR domain always preserves the end-to-end integrity of traffic traversing it. No extension header is manipulated, inserted or removed from an inner transported packet. The packet leaving the SR domain is exactly the same (except for the hop-limit update) as the packet entering the SR domain.¶
The SR domain is designed with link MTU sufficiently greater than the MTU at the ingress edge of the SR domain.¶
The following deployments are as of November 2019.¶
Six operators have publicly reported SRv6 deployments with commercial traffic supported by linerate hardware. Each deployment follows the network design and SRH add/remove behavior described in this document.¶
Further information can be found in [I-D.matsushima-spring-srv6-deployment-status]¶
Eighteen unique implementations of SRv6 are available from multiple vendors and open source initiatives that support the SRH add/remove behavior described in this document:¶
Further information can be found in [I-D.matsushima-spring-srv6-deployment-status]¶
An SR Domain is defined in [RFC8402].¶
Section 5.2 of [I-D.ietf-6man-segment-routing-header] further describes the SR domain as a single system with delegation among components. It states:¶
In other words, all packets within the SR domain have a source and destination address within the SR Domain.¶
The SR domain is secured as per Section 5.1 of [I-D.ietf-6man-segment-routing-header] and no external packet can enter the domain with a destination address equal to a segment of the domain.¶
In other words, no node outside the SR domain may send packets to, nor make direct use of, segments within the SR domain.¶
The following abstract illustration shows the SR Domain, how traffic is encapsulated when traversing the SR domain, and (in subequent sections) how an SRH is inserted and processed for a packet traversing the SR domain. It is representative of all deployments in Section 2.1.¶
+ * * * * * * * * * * * * * * * * * * * * +
* *
[1]----[3]--------[5]----------------[6]---------[4]---[2]
* | | *
| |
* | | *
[7]----------------[8]
* *
+ * * * * * * * SR Domain * * * * * * * +
Since all inter domain packets are encapsulated for the part of the packet journey that is within the SR Domain, a packet sent from 1 and destined to 2 is encapsulated in an outer IP v6 header between nodes 3 and 4.¶
Without revealing the specifics of each deployment, the following well-known technique can be used:¶
Vendors and operators have automated the process of locator selection, the details of which are outside the scope of this document.¶
The security measures defined in [I-D.ietf-6man-segment-routing-header] Section 5.1 are applied.¶
Protection level 1: filter external traffic entering the SR domain. For example, node 4 (on its interface from node 2) applies an ingress ACL that drops any packet with DA within the PPPP:PPPB:BB00::/40 block.¶
Protection level 2: filter internal traffic. For example, node 4 (on its interface from node 6) applies an ingress ACL that drops any packet with DA in PPPP:PPPB:BB00:0004::/64 block if SA is not within the block PPPP:PPP0::/28¶
Vendors and operators have automated the application of these protection levels, the details of which are outside the scope of this document.¶
The deployments, Section 2.1, leverage the extensively used practice of ensuring an MTU within the domain is bigger than the MTU on the external links of the domain.¶
More specifically, the MTU difference (MTU-Delta) is designed to be larger than the maximum encapsulation overhead deemed required by the deployment.¶
The exact number is operator specific and is outside the scope of this document. Some indications on how to plan this are provided in the following sections.¶
Any packet exceeding the MTU of a link generates an IPv6 ICMP error message "packet too big" back to the source of the packet.¶
The deployments involve the creation of commercial SRv6-based VPN traffic as described in [I-D.ietf-bess-srv6-services].¶
The salient point to note is that no SRH needs to be inserted to realize an SRv6 VPN service.¶
The ingress PE encapsulates the inner packet in an outer header and sets the outer DA to the END.DT/DX SID signaled by the egress PE.¶
MTU-Delta must be >= 40 bytes to allow for the outer IPv6 encapsulation without fragmentation.¶
The deployments involve the delivery of sub-50msec TILFA protection to the commercial SRv6-based VPN traffic transported by the operator network [I-D.ietf-rtgwg-segment-routing-ti-lfa].¶
In these deployments, when a failure is detected, the Point of Local Repair (PLR) inserts an SRH implementing the precomputed TILFA backup path.¶
The following salient points are discussed:¶
When an SRH is inserted by an intermediate node it walks the IPv6 header chain to the first header after the IPv6 header and inserts the SRH prior to that header.¶
+---------------+------------
| IPv6 header | IPv4 header
| VPN Service |
| Next Header = |
| IPv4 |
+---------------+------------
^-SRH insertion here
An SR Policy headend within the SR domain inserts an SRH as follows:¶
Increase the IPv6 header payload length field by the length in bytes of the inserted SRH.¶
If the resulting payload length exceeds 2^16 bytes generate an ICMP "Packet To Big" error message to the source with an MTU of 2^16 minus the length in bytes of the SRH and discard the packet.¶
Note: this does not occur in reported deployments given the MTU design constraint.¶
The TILFA protection service is essentially a transparent service: it seeks to make the loss of a link, node or SRLG invisible to the transport service. It is also a very transient service as it only lasts a few hundreds of msec while the IGP converges.¶
Consistent with this transparent service definition, the deployments leverage a TILFA computation that ensures that the penultimate SID of the inserted SRH is of PSP flavor.¶
The vendors reporting the listed deployments have collectively deployed TILFA in tens of SR-MPLS networks, in 6 SRv6 networks and have simulated their SRv6 algorithm in tens of collected real topologies. The inferred experience is that the probability that a TILFA backup path requires more than 2 SRv6 SIDs is very rare.¶
MTU-Delta must be >= 80 bytes.¶
The maximum encapsulation size of any node within the SR domain is limited to a specific value, this maximum is used to calculate the maximum link MTU of interfaces ingress to the SR domain.¶
Section 6 describes the method of securing the SR domain in the deployments listed.¶
All security considerations discussed in [I-D.ietf-6man-segment-routing-header] are equally applicable when an SRH insertion is performed.¶
This document doesn't introduce any IANA request.¶
The authors would like to thank the following for their contributions: Robert Raszuk, Stefano Previdi, Stefano Salsano, Antonio Cianfrani, David Lebrun, Olivier Bonaventure, Prem Jonnalagadda, Milad Sharif, Hani Elmalky, Ahmed Abdelsalam, Arthi Ayyangar, Dirk Steinberg, Wim Henderickx.¶