Router Protocol Process Migration via Graceful Restart
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
State-of-the-art IP routers experience traffic loss during routing convergence due to protocol failures, and existing solutions for resiliency and scalability are complex and costly, particularly in handling processor failures and migrating protocol processes.
Innovation Solution
A method for migrating routing protocol processes from one Route Processor to another using graceful restart procedures, which includes terminating protocol processes, adding filter rules to divert packets, and restarting processes on a new processor without requiring changes to the network protocol stack, ensuring minimal packet loss and maintaining network connectivity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If protocol processes are terminated on the first route processor to enable migration, then process migration capability is improved, but packet loss occurs during routing convergence
Solution Approach 1:
The forwarding engine is configured in advance to divert packets destined for the terminating process to the new route processor. This preliminary action ensures that when the process migration occurs, packet forwarding continues uninterrupted without requiring changes to the network protocol stack, thus resolving the contradiction between enabling migration and preventing packet loss
Solution Approach 2:
The forwarding engine acts as an intermediary between the terminating route processor and the new route processor. It receives packets destined for the old process and redirects them to the new process, enabling seamless migration while maintaining network reliability and preventing packet loss during the transition
2Reliability
If hot-standby backup route processors are implemented, then resiliency against node failures is improved, but complexity in synchronizing state information increases
Solution Approach 1:
The patent extracts the state synchronization complexity from the backup route processor configuration. Instead of requiring continuous synchronization between primary and backup RPs, the system uses the forwarding engine to redirect packets to the new RP during migration, eliminating the need for complex state synchronization while maintaining resiliency
Solution Approach 2:
The system copies the operational state and packet flow information to the new route processor during migration. The forwarding engine replicates the necessary packet redirection logic, allowing the backup RP to take over without requiring complex real-time synchronization of all state information, thus reducing overall system complexity
3Productivity
If processor clusters are used to handle scalability needs, then processing capacity is improved, but implementation complexity and cost increase
Solution Approach 1:
The forwarding engine is designed with multi-functionality to handle both normal packet forwarding and process migration redirection. This universal design allows the same hardware infrastructure to support scalability without requiring complex additional components, reducing implementation complexity while maintaining processing capacity
Solution Approach 2:
The system enables self-service migration where the forwarding engine automatically redirects packets to the new route processor without requiring manual intervention or complex coordination between multiple processors. This self-service capability simplifies the migration process and reduces the operational complexity of processor clusters
Data Source
AI summary
A router and a method for migrating routing protocol processes or Virtual Routers (VRs) from one Route Processor (RP) to another using graceful restart procedures for maintaining packet flow to the router and assisting the router to obtain restart information. The router includes first and second RPs and a forwarding engine for forwarding packets to neighbor routers in the network. When the routing protocol process is terminated on the first RP, the neighbor routers detect a router failure and initiate the graceful restart procedures. A filter rule in the forwarding engine causes it to forward packets to the new routing protocol process on the second RP. The new process learns the network topology from the neighbor routers, and the migration is completed without packet loss and without requiring complex changes to the network protocol stack.