FIB Entry Optimizer for Selective Route Download Traffic Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network infrastructure faces inefficiencies in packet forwarding due to the computational cost of determining routes, as the routing information base (RIB) is costly to compute, while the forwarding information base (FIB) may not include entries for all possible IP prefixes, leading to suboptimal routing decisions.
Innovation Solution
A method and system where a FIB entry optimizer ranks entries of the RIB and FIB based on access information and estimated access rates, deprogramming less important entries and programming more important ones into the FIB to optimize routing, thereby improving packet forwarding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If the FIB includes entries for all possible IP prefixes to improve routing accuracy, then the manufacturing precision of routing decisions is improved, but the device complexity and computational cost increase significantly
Solution Approach 1:
The patent segments the routing decision process into two stages: (1) FIB lookup for common prefixes with pre-computed next-hops, and (2) RIB computation for prefixes not in the FIB. This segmentation allows the system to maintain high routing accuracy while reducing FIB complexity by only including frequently accessed entries.
Solution Approach 2:
The patent performs preliminary computation of next-hops for common prefixes during off-peak times and pre-loads them into the FIB. This preliminary action reduces the computational burden during peak routing operations, as most lookups can be satisfied by pre-computed entries without requiring full RIB computation.
2Productivity
If the FIB is maintained with comprehensive entries to improve routing performance, then the productivity of packet forwarding is improved, but the loss of time for FIB updates and maintenance increases
Solution Approach 1:
The patent implements periodic FIB updates rather than continuous synchronization with the RIB. The FIB is updated at scheduled intervals or based on traffic pattern changes, allowing the system to maintain high forwarding throughput while periodically refreshing routing information without constant disruption.
Solution Approach 2:
The patent applies partial updates to the FIB, updating only the entries that have changed in the RIB rather than performing full FIB regeneration. This partial action reduces the time loss associated with FIB maintenance while still maintaining routing accuracy for affected prefixes.
3Reliability
If the RIB is fully computed to ensure complete routing information, then the reliability of routing decisions is improved, but the use of energy and computational resources increases
Solution Approach 1:
The patent extracts only the essential next-hop information from the full RIB computation and stores it in the FIB for common prefixes. This extraction approach maintains routing reliability by preserving critical forwarding information while reducing the energy cost of maintaining complete routing state in the fast-path forwarding table.
Solution Approach 2:
The patent introduces the FIB as an intermediary layer between the RIB and the packet forwarding plane. The FIB acts as a cache that stores pre-computed next-hops for frequently accessed prefixes, reducing the need for full RIB computation during normal operation while maintaining routing reliability through periodic updates.
Data Source
AI summary
A network device includes a forwarding information base (FIB). The FIB includes a first number of entries and a default entry. The network device includes a routing information base that includes a second number of entries. The network device includes a FIB entry optimizer that ranks a first portion of the second number of entries based on access information of the first number of entries; ranks a second portion of the second number of entries based on access information of the default entry; and updates at least one entry of the FIB based on the ranks of the first portion of the second number of entries and the ranks of the second portion of the second number of entries. The first number of entries is less than the second number of entries.