Network Device Conditional Flooding for Topology Change Convergence
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network convergence times are prolonged due to the need to flush and relearn forwarding database entries in response to topology changes, which is inefficient and affects failover times, especially when only a single link fails, causing unnecessary flushing of unaffected ports and links.
Innovation Solution
Implementing conditional packet flooding by clearing an age-bit array to invalidate entries and overriding initial flooding, allowing packets to be forwarded conditionally based on the validity of forwarding entries, reducing the need for extensive flushing and relearning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If forwarding database entries are flushed in response to topology change, then forwarding correctness is ensured, but convergence time increases
Solution Approach 1:
The patent segments the forwarding database into multiple tables (first forwarding table and second forwarding table) and divides the flushing operation into partial flushes. Instead of flushing all entries simultaneously, only specific portions are flushed at different times, allowing the system to maintain forwarding correctness while reducing the overall convergence time by avoiding unnecessary flush operations.
Solution Approach 2:
The patent performs preliminary actions by pre-flushing specific portions of the forwarding database before the actual topology change occurs or by proactively invalidating specific entries based on topology change predictions. This preliminary partial flushing reduces the workload during the actual convergence event, thereby reducing overall convergence time while ensuring forwarding correctness.
2Measurement precision
If all forwarding entries are flushed, then forwarding accuracy is maintained, but processing overhead increases
Solution Approach 1:
The patent applies local quality by treating different portions of the forwarding database differently. Instead of uniformly flushing all entries, the system identifies and flushes only the specific portions that are affected by topology changes. This localized approach maintains forwarding accuracy for affected entries while avoiding unnecessary processing overhead for unaffected entries.
Solution Approach 2:
The patent implements partial action by performing only the necessary flushing operations rather than flushing all forwarding entries. By identifying and flushing only the specific portions that need updating, the system maintains forwarding accuracy while significantly reducing processing overhead compared to a complete flush of the entire forwarding database.
3Speed
If forwarding database is organized as hash table, then lookup efficiency improves, but flushing complexity increases
Solution Approach 1:
The patent segments the hash table-based forwarding database into multiple manageable tables or portions. This segmentation allows the system to maintain the lookup efficiency of hash tables while simplifying flushing operations by working with smaller, divided portions rather than the entire large hash table, thereby reducing flushing complexity.
Data Source
Figure 1A
Figure 1B
Figure 2
AI summary
In response to a network topology change, packets are initially flooded on ports of a network device. In addition, a bit array is cleared in response to the topology change. Each bit in the bit array is associated with a particular forwarding entry on the network device. In connection with the clearing of the bit array, the flooding of packets on ports of the network device is made conditional, reducing failover time of the network.