Source Packet Routing in Networking (SPRING) or Segment Routing

Support for S-BFD over SRv6-TE paths with Classic and Micro SIDs (PTX10002-36QDD, PTX10004, PTX10008, and PTX12008)—You can validate and control Segment Routing for IPv6—Traffic Engineering (SRv6-TE) path liveness with Seamless-BFD (S-BFD) for Classic and Micro Segment Identifiers (SIDs). S-BFD helps improve path convergence and reliability by ensuring that a SRv6-TE path is usable only when its S-BFD session is up. S-BFD sessions run in distributed mode and support non-stop routing (NSR) and GRES.

To interoperate with responders that expect the IPv6 local-host destination, use the set protocols source-packet-routing source-routing-path lsp-path-name primary segment-list-name bfd-liveness-detection sbfd destination-ipv6-local-host configuration statement.

You can display SRv6 S-BFD session details with the show spring-traffic-engineering sbfd command.

[See S-BFD for SRv6 TE Paths.]

Color-based forwarding (CBF) for SRv6-TE (PTX10002-36QDD, PTX10004, PTX10008, PTX10016, and PTX12008)—We extend the existing color-based forwarding (CBF) functionality to SRv6 traffic engineering enabling you to adapt to complex network demands. Use CBF for SRv6-TE to steer traffic across multiple transport tunnels based on class of service (CoS). This approach improves route selection, next-hop resolution, and service quality. Resolver enhancements support SRv6 Segment Identifiers (SID) across multipath routes, and the preserve-next-hop-hierarchy configuration prevents misrouting.

[See preserve-nexthop-hierarchy (SRv6-TE).]

Support for UHP in IS-IS SR-MPLS (ACX7020, ACX7100, ACX7332, ACX7348, ACX7509,ACX7024, PTX10002-36QDD, PTX10004, PTX10008, PTX10016, and PTX12008) —Use Ultimate Hop Popping (UHP) with IS-IS or OSPF so the egress provider edge (PE) can process its own node SID. ISIS advertises a node SID with the P flag set and E flag unset. In controller-driven segment routing traffic engineering (SR-TE) the controller inserts the egress PE node SID beneath the SR-TE binding SID. If the Binding SID route fails on the penultimate hop, the egress PE might see its own node SID as the top label instead of penultimate hop popping (PHP). With the P flag set, the PE expects UHP and processes its MPLS label. Include the ultimate-hop-popping statement at the [edit protocols isis source-packet-routing] hierarchy level.

[See ultimate-hop-popping.]

SRv6-TE route resolution over BGP without IGP (PTX10002-36QDD, PTX10004, PTX10008, and PTX10016)—An SRv6 tunnel consists of paths with segment identifiers (SIDs) over IGP that steer traffic to a traffic-engineering (TE) path. If IGP is not available, configure these SIDs statically and advertise them through External BGP (EBGP). This feature is supported on both classic and micro SRv6 SIDs

Networks that deploy network orchestrator to steer transit traffic onto a TE path and advertise these transit prefixes using BGP color community don't have a service SID. In this case, the last SID must not be removed, and the ingress SRv6 TE tunnel acts as a transit tunnel to forward transit traffic with SRv6 encapsulation.

Include the no-remove-srv6-last-sid statement at the [edit protocols source-packet-routing] hierarchy level and the use-ingress-routes-as-transit statement at the [edit protocols source-packet-routing srv6] hierarchy level.

[See no-remove-srv6-last-sid.]

Delay normalization for OSPF Flexible Algorithm metrics and advertisements across IGP instances (ACX7020-AC, ACX7020-AC-C, ACX7020-DC, ACX7020-DC-C, ACX7100-32C, ACX7100-48L, ACX7332, ACX7348, ACX7509, ACX7024, ACX7024X, PTX10001-36MR, PTX10002-36QDD, PTX10003, PTX10004, PTX10008, PTX10016, PTX12008)—Use delay normalization to compute and advertise a normalized delay metric for Flexible Algorithm, to improve path-selection consistency across all IGP instances. The device normalizes each received delay, compares each value with the previously saved normalized value, and triggers link-state advertisement (LSA) generation when the values differ.

Delay normalization is disabled by default. To enable and configure delay normalization, use the normalize interval offset statement at the [edit protocols ospf area interface delay-measurement] hierarchy level.

[See delay-measurement (Protocols OSPF) and How to Configure Flexible Algorithms in OSPF for Segment Routing Traffic Engineering.]

Selectively control per-prefix backup paths with OSPF import policy (ACX7020-AC, ACX7020-AC-C, ACX7020-DC, ACX7020-DC-C, ACX7100-32C, ACX7100-48L, ACX7332, ACX7348, ACX7509, ACX7024, ACX7024X, PTX10001-36MR, PTX10002-36QDD, PTX10003, PTX10004, PTX10008, PTX10016, PTX12008)—You can selectively enable backup paths for specific prefixes to optimize redundancy and resource utilization. By default, a configured backup path applies to all prefixes. To exclude specific prefixes or ranges, create an OSPF import policy and configure the no-backup option in the then clause of the policy to suppress backup path installation for matching routes. You can reserve backup protection for critical prefixes while preventing unnecessary backups for others.

[See Understanding Backup Selection Policy for OSPF Protocol.]

Configure flex-algorithm-preference statement at the [edit protocols ospf] hierarchy level to prioritize desired routes and improve traffic engineering across IP and MPLS domains.

Policy-based redistribution of OSPF prefix SIDs across IGP instances (ACX7020-AC, ACX7020-AC-C, ACX7020-DC, ACX7020-DC-C, ACX7100-32C, ACX7100-48L, ACX7332, ACX7348, ACX7509, ACX7024, ACX7024X, PTX10001-36MR, PTX10002-36QDD, PTX10003, PTX10004, PTX10008, PTX10016, PTX12008)—You can redistribute Segment Routing (SR) prefix-SIDs across OSPF IGP instances using route policy without explicitly specifying a prefix-segment index. This feature standardizes SR labels across instances and improves operational efficiency. Configure a policy with the from prefix-segment statement to match routes carrying prefix-segment information. In the then clause, use prefix-segment redistribute to inherit segment information from the matched route. We also support stitching mpls.0 routes to enable interoperability between different IGP instances.

Non-router-ID endpoints as SR-TE destinations (ACX7020-AC, ACX7020-AC-C, ACX7020-DC, ACX7020-DC-C, ACX7024, ACX7024X, ACX7100-32C, ACX7100-48L, ACX7332, ACX7348, ACX7509, PTX10001-36MR, PTX10001-36MR-K, PTX10002-36QDD, PTX10002-60MR, PTX10003, PTX10004, PTX10008, PTX10016, and PTX12008)—Use non-router-ID endpoints as destinations in Segment Routing—Traffic Engineering (SR-TE) policies for Segment Routing for MPLS (SR-MPLS). Traditionally, these policies use router IDs, but you can specify anycast addresses to enhance redundancy and load balancing in SR-MPLS networks. Use IPv4 and IPv6 anycast addresses as IGP-learned destinations with or without Segment Identifier (SID) stack compression. These anycast addresses are not redistributed (R-bit set). Use them as the to address for SR-TE policies with associated compute profiles..

[See Non-Router-ID Endpoints in Segment Routing Traffic Engineering.]

Static SID configuration in SRv6 Manager (PTX10002-36QDD, PTX10004, PTX10008, PTX10016, and PTX12008)—Configure SRv6 classic, micro node, adjacency SIDs, along with classic END and END-X SIDs, and install them in the routing table without using interior gateway protocols (IGPs) such as IS-IS. Advertise these static routes through a BGP export policy for path computation. This configuration enables controllers to receive static node and adjacency SIDs over BGP and compute SR-TE paths across a domain that does not use IS-IS.

[See Understanding SRv6 Static Segment Identifier.]