FDB Segmentation for Rapid Network Failure Recovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In networks with multiple L2 switches, when a failure occurs, the time-consuming process of disabling and rebuilding the FDB (Filtering Database) entries leads to increased network traffic and operational disruptions due to flooding before address learning is complete.
Innovation Solution
A communication apparatus and method that updates the FDB by setting a failure state number, allowing for partial disabling of entries and prioritizing new learning, reducing the need for full FDB flush and minimizing processing time, even with increased network size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If full FDB flush is performed when network failure occurs, then reliability is improved by ensuring accurate forwarding paths, but time consumption increases and network traffic increases due to flooding
Solution Approach 1:
The patent segments the FDB into valid entries (with failure state number matching current failure state) and invalid entries (with mismatched failure state numbers). Instead of flushing the entire FDB, only invalid entries are removed, while valid entries are preserved. This segmentation approach maintains forwarding path accuracy for affected paths while preserving useful routing information for unaffected paths, thereby reducing reconstruction time and minimizing flooding.
2Reliability
If full FDB flush is performed, then forwarding correctness is ensured, but network traffic increases due to flooding during reconstruction
Solution Approach 1:
The FDB is segmented into valid and invalid entries based on failure state numbers. Valid entries (where failure state number matches current failure state) are preserved and used for correct forwarding decisions. Invalid entries are removed. This segmentation ensures forwarding correctness for affected paths while avoiding the need to flood the entire network during reconstruction, thereby reducing unnecessary network traffic.
Solution Approach 2:
The patent performs preliminary action by maintaining failure state numbers in the FDB entries and using them to identify which entries need to be removed. When a failure occurs, the system already has the failure state number information stored in valid FDB entries, allowing it to quickly identify and remove only the necessary invalid entries without needing to flood the network or perform a complete FDB flush, thus reducing network traffic.
3Productivity
If FDB reconstruction is performed quickly, then operational disruptions are reduced, but accuracy of address learning may be compromised
Solution Approach 1:
The patent segments FDB entries into valid and invalid categories based on failure state numbers. Valid entries (with matching failure state numbers) are preserved and used for accurate forwarding decisions. Invalid entries are removed. This segmentation enables quick reconstruction by processing only the necessary invalid entries rather than the entire FDB, while maintaining accuracy through the preservation of valid entries that contain correct routing information.
Solution Approach 2:
The patent changes the parameter of failure state number storage in FDB entries from a static field to a dynamic indicator of validity. By using failure state numbers as a validity indicator, the system can quickly determine which entries are valid and which need removal. This parameter change enables both fast reconstruction (by targeting only invalid entries) and high accuracy (by preserving valid entries with correct routing information).
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
The present invention provides a communication apparatus that operates as a node of a network. The communication apparatus includes a port-state managing unit 340 configured to update, every time failure occurrence in the network is detected, state information indicating a state of the network, a database updating unit configured to update, based on a transmission source address and a reception port of a received frame and the state information managed by the port-state managing unit 340, a database for determining an output port when a frame in which the transmission source address is a destination address is received, and an output-destination determining unit configured to determine an output port of the received frame by searching through the database using, as search keys, the destination address of the received frame and the state information managed by the port-state managing unit 340.