Redundant Packet Forwarding Engines for Accurate Switchover Statistics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In redundant routing systems, maintaining accurate traffic statistics during a switchover between primary and backup routing engines is challenging due to discrepancies in tracking and the resulting high steady-state load, which affects router performance.
Innovation Solution
A method where the backup routing engine requests and stores traffic statistics from the primary engine as residual data, resets counters, and combines this with duplicated traffic to maintain continuity during a switchover, minimizing discrepancies and reducing load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the backup routing engine independently tracks traffic statistics, then reliability is improved through redundancy, but measurement precision deteriorates due to discrepancies in tracking between primary and backup engines
Solution Approach 1:
The backup routing engine receives and stores a copy of traffic statistics from the primary engine before switchover occurs. This preliminary action ensures that when switchover happens, the backup engine already has accurate residual statistics to combine with its own tracking data, eliminating discrepancies without requiring independent tracking during normal operation.
2Adaptability or versatility
If the backup routing engine independently tracks traffic statistics, then adaptability is improved for failover, but loss of information occurs due to discrepancies in traffic data between primary and backup engines
Solution Approach 1:
The backup routing engine combines two sources of traffic statistics: residual statistics received from the primary engine and statistics independently tracked during backup operation. By merging these data sources and using their average as weighting, the system maintains continuous accurate traffic statistics throughout the failover process, preventing information loss.
3Measurement precision
If traffic statistics are synchronized frequently between primary and backup engines, then measurement precision is improved, but productivity deteriorates due to high steady-state load
Solution Approach 1:
Instead of continuous synchronization, the system uses periodic action by having the backup engine receive traffic statistics from the primary engine at specific intervals (e.g., periodically or at switchover events). This reduces the steady-state load on the system while maintaining sufficient accuracy for failover requirements.
Data Source
AI summary
In some implementations, a first processing component of a network device may receive first traffic data obtained by a second processing component of the network device. The first processing component may store the first traffic data as residual statistics. The first processing component may obtain second traffic data associated with a copy of a traffic stream processed by the first processing component based on storing the first traffic data as the residual statistics. The first processing component may perform a switchover from the second processing component to the first processing component. The first processing component may determine current traffic data based on the residual statistics and the second traffic data. The current traffic data may be determined based on performing the switchover from the second processing component to the first processing component.