Segment Routing Branch SID for Label Depth and Inter-Domain Protection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
MPLS and Segment Routing-based networks face challenges in inter-domain traffic protection due to label stacking limitations, which constrain the number of labels that can be imposed, leading to insufficient protection and non-compliance with service level agreements, especially in heterogeneous networks.
Innovation Solution
The introduction of a Branch Segment Identifier (SID) that combines the functionalities of an anycast SID and a binding SID, allowing for optimal label depth and protection across domains by allocating and flooding it on dual-homed border nodes, enabling seamless routing and failure recovery.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If label stacking is used to provide protection paths in inter-domain networks, then reliability is improved, but label stack depth is exceeded due to hardware limitations
Solution Approach 1:
The patent segments the protection mechanism by introducing a specialized protection SID that operates independently from the data path labels. This segmentation allows the protection path to be established without adding multiple labels to the data packet, thus maintaining label stack depth within hardware limits while still providing redundant path capability
Solution Approach 2:
The patent introduces a protection SID as an intermediary element that mediates between the data plane and the control plane. This intermediary enables fast reroute to protection paths by serving as a single-label shortcut that bypasses the need for deep label stacks, while still providing comprehensive protection across domain boundaries
2Reliability
If end-to-end signaling is used to compute protected paths, then reliability is improved, but flexibility is reduced and service level agreements cannot be guaranteed
Solution Approach 1:
The patent applies preliminary action by pre-computing and advertising protection paths through anycast SID before failures occur. This allows border routers to have protection paths ready in advance, enabling fast convergence that meets service level agreements without requiring complex end-to-end signaling during actual failure events
Solution Approach 2:
The patent introduces dynamics by allowing the protection mechanism to adapt to different failure scenarios through anycast SID. The system can dynamically select among multiple border routers based on real-time conditions while maintaining pre-established protection paths, providing both reliability and adaptability to meet varying service level requirements
3Adaptability or versatility
If multiple labels are imposed on packets for inter-domain routing, then routing flexibility is improved, but hardware constraints are exceeded
Solution Approach 1:
The patent extracts the protection path selection function from the data packet labels and places it in the control plane through anycast SID advertising. This extraction removes the need to carry multiple protection labels in the data plane, reducing label stack depth while maintaining routing flexibility through the extracted anycast SID mechanism
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A source node (11) in a first domain (6) in a multi-domain network (5) includes a plurality of ports; and switching circuitry configured to forward packets between the plurality of ports, wherein the source node is configured to, for a packet destined for a destination node (33) in another domain in the multi-domain network (5), add (81) a list of Segment Identifiers (SIDs) to the packet for loose source-based routing from the source node (11), through the first domain (6), and to the another domain, and forward (82) the packet via a port of the plurality of ports and with the list of SIDs, wherein the list of SIDs includes at least one SID representing both an anycast SID and a binding SID. The SID is associated with an anycast group that includes a plurality of nodes (21, 26), and with a binding segment in the second domain (7), the binding segment includes a different SID list for each node of the plurality of nodes (21, 26).