FIB Caching System Non-Overlapping Prefixes Cache Hiding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The rapid growth of the Routing Information Base (RIB) leads to increased size and cost of Forwarding Information Base (FIB) memory in routers, with FIB caching facing issues like cache-hiding problems and high miss ratios due to network traffic locality and prefix changes.
Innovation Solution
A FIB caching system that selects and generates minimal non-overlapping prefixes for the cache, using a space-optimized tree structure and systematic cache management to handle misses and updates, avoiding cache-hiding issues and achieving high hit ratios with reduced memory usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If FIB caching is implemented to reduce memory cost and improve lookup speed, then memory cost and lookup speed are improved, but cache-hiding problems occur where packets are forwarded to wrong next hops due to longest-prefix match mismatches between cache and full FIB
Solution Approach 1:
The patent segments the FIB lookup process into two distinct phases: cache lookup for speed and full FIB lookup for accuracy. The segmentation separates the fast-path caching mechanism from the slow-path verification mechanism, allowing the system to benefit from both high-speed caching and accurate longest-prefix matching without the cache-hiding problem compromising forwarding correctness
Solution Approach 2:
The patent introduces an intermediary verification mechanism that acts as a mediator between the cache and the forwarding decision. When a cache hit occurs, the system uses an intermediary check to verify whether the cached entry is still valid and consistent with the full FIB before finalizing the forwarding decision, thus preventing cache-hiding problems from causing incorrect routing
2Productivity
If FIB cache size is increased to improve hit ratio, then hit ratio is improved, but memory cost and power consumption increase
Solution Approach 1:
The patent changes the parameter of cache entry structure by implementing a compact representation that stores only essential forwarding information in the cache rather than full FIB entries. This parameter change allows the cache to achieve high hit ratios with significantly reduced memory capacity, as each cache entry consumes less space while still providing accurate forwarding decisions
Solution Approach 2:
The patent applies partial action by implementing a two-stage lookup process where the cache handles the majority of lookups (partial action) and the full FIB is consulted only when necessary. This approach achieves high productivity through the cache while avoiding the need to allocate memory for the entire FIB in the fast path, thus optimizing the balance between hit ratio and memory capacity
3Productivity
If FIB cache is used to speed up forwarding, then forwarding speed is improved, but cache replacement and updates become complex leading to cache-hiding problems
Solution Approach 1:
The patent extracts the complexity of cache management by separating the cache update logic from the forwarding path. Cache replacement and updates are handled by a dedicated management component that operates independently from the high-speed forwarding engine, thus simplifying the forwarding path while still maintaining accurate cache contents and reducing cache-hiding problems
Data Source
AI summary
A Forwarding Information Base (FIB) caching system and scheme that selects and generates a minimal number of non-overlapping prefixes for the FIB cache to improve computer-based network communications and operations efficiency, accuracy and speed. Because the cached prefixes do not cover any longer prefixes in the full FIB, the present invention thus avoids cache-hiding issues. The system further comprises a FIB caching updating component that systematically handles cache misses, cache replacement, and routing updates.