SR-MPLS Flexible Algorithm

Current network infrastructures, mostly for 5G services, employ differentiated traffic handling given constraints such as delay requirements, bandwidth, or path availability. In most cases, shortest‐path routing does not always evenly load the link traffic, causing congestion on critical links. For that reason, RFC 9350 introduced Flexible Algorithm (Flex‐Algo) via which operators could proactively “slice” the network into multiple planes, each with their own routing rules and metrics. In our testing, we considered using three Flex‐Algo definitions—FA128, FA129, and FA130 to show different routing requirements. The FA128 prioritized paths with lower latency by using static or dynamic delay measurements to assess the path. The FA129 optimized traffic engineering (TE) metrics to balance link utilization among the available multiple links. Finally, FA130 leveraged link affinities (or “colors”), a constraint that excludes certain links in the path computation. These enhancement measures are possible only because each node or router advertises its algorithm capabilities through IGP. In IS-IS, these are presented in additional sub-TLVs and/or Application-Specific Link Attribute(ASLA) and assigned prefix SIDs mapped to the given Algorithm ID, making it possible to construct a novel logical topology given the constraints of the router. As defined in all three conditions, the routers successfully created paths, and paths were correctly installed on MPLS forwarding tables, allowing end‐to‐end traffic to be passed. We conclude that our experiment illustrates the fact that the use of the Flex‐Algo gives operators the ability to provide a good performance and resilience service class and allows the simplification of configuration in SR‐MPLS environments.