Two-Phase Fast Reroute for IP Network Restoration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current IP network restoration methods, such as IP reroute and MPLS Fast Reroute, face challenges in achieving sub-100 ms restoration while maintaining efficient bandwidth utilization and latency behavior, especially in scenarios involving single or multiple failures, and are not optimized for resource utilization and cost minimization.
Innovation Solution
A two-phase fast reroute system with optimized traffic engineering that uses pre-computed next-hop or next-next-hop backup tunnels and dynamic link weights to reroute traffic quickly, followed by optimized end-to-end path rerouting using RSVP-TE convergence, ensuring sub-100 ms restoration with minimal traffic loss and efficient bandwidth utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If MPLS Fast Reroute is used for sub-100 ms restoration, then restoration speed is improved, but bandwidth utilization efficiency deteriorates
Solution Approach 1:
The restoration process is segmented into two distinct phases: Phase 1 uses pre-computed backup tunnels for immediate sub-100 ms restoration, while Phase 2 performs optimized rerouting for long-term efficiency. This segmentation allows the system to achieve fast restoration without permanently sacrificing bandwidth utilization, as the suboptimal backup paths are only used temporarily during the transition phase.
Solution Approach 2:
Backup tunnels are pre-computed and pre-configured before failures occur, enabling immediate traffic switching upon failure detection. The pre-computation includes establishing backup Label Switched Paths and configuring forwarding tables in advance, which eliminates computation delays during actual restoration events and achieves sub-100 ms switching time.
2Loss of energy
If IP reroute with IGP convergence is used, then bandwidth utilization is improved, but restoration time deteriorates
Solution Approach 1:
The system performs preliminary computation of backup tunnels and pre-configures restoration paths before failures occur. By pre-computing backup Label Switched Paths and configuring forwarding tables in advance, the system eliminates the need for real-time path computation during failures, achieving sub-100 ms restoration without sacrificing long-term bandwidth optimization.
3Speed
If backup paths follow shortest paths in MPLS FRR, then restoration speed is improved, but capacity utilization deteriorates
Solution Approach 1:
The system dynamically transitions between different routing phases: Phase 1 uses pre-computed backup tunnels (which may follow shortest paths for speed) immediately upon failure, while Phase 2 performs optimized rerouting that considers long-term capacity utilization. This dynamic approach allows the system to prioritize speed when needed and efficiency when stable, resolving the contradiction between restoration speed and capacity utilization.
4Loss of time
If sub-100 ms restoration is achieved using pre-computed backup tunnels, then restoration time is improved, but overall network cost deteriorates
Solution Approach 1:
The restoration process is segmented into two phases with different optimization goals. Phase 1 uses pre-computed backup tunnels for fast restoration, accepting higher temporary resource usage. Phase 2 performs optimized rerouting to reduce long-term resource consumption. This segmentation allows the system to achieve fast restoration without permanently over-provisioning network capacity, as the suboptimal resource usage is temporary and followed by optimization.
Data Source
AI summary
Systems and methods are described for restoring IP traffic flows routed in an IP network in less than 100 ms after the IP network sustains a network failure. The systems and methods are a two-phase fast reroute that uses distributed optimized traffic engineering during backup tunnel restoration and end-to-end tunnel restoration that maximizes sharing among all independent failure scenarios and minimizes the total capacity, total cost, or a linear combination of the two. For defined failure condition scenarios, restoration traffic is routed using pre-computed backup tunnels using a constrained shortest path first method where link weights along the path are chosen dynamically and depend on available and total capacity of the link, latency, and other cost measures such as IP port costs. During the capacity allocation phase, the method reuses capacity already allocated for other independent failure scenarios as much as possible but also adds capacity, if necessary. When an actual IP network failure occurs, the backup tunnels are used to immediately restore service to affected IP network traffic flows. In parallel, end-to-end tunnels corresponding to each affected traffic flow are rerouted and once the rerouting process is complete, traffic is switched over from the old end-to-end tunnel routes (using backup tunnels) to new end-to-end tunnel routes without using backup tunnels.