Hash Table Prefix Length Conversion for Routing Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing routing tables in networking devices are inefficient due to their large size and the resource-intensive nature of ternary content addressable memory (TCAM) usage, which consumes significant power and memory, especially during IP address prefix lookup operations.
Innovation Solution
The solution involves breaking down a large routing table into smaller hash tables of unified prefix lengths by converting smaller prefix lengths into equivalent lengths, allowing for the creation of a single unified hash table that reduces the number of tables and prefix lengths, and storing this hash table in RAM instead of TCAM to optimize resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a large routing table is stored in TCAM to perform IP address prefix lookup operations, then the routing functionality is maintained, but power consumption and memory usage increase significantly
Solution Approach 1:
The patent segments the large routing table into multiple smaller hash tables organized by prefix length ranges. Each hash table stores a specific subset of routing entries, allowing the system to perform lookups in smaller, more efficient units rather than searching the entire routing table, thereby reducing power consumption while maintaining routing functionality
Solution Approach 2:
The patent replaces expensive, high-power TCAM memory with cheaper, lower-power RAM memory for storing routing table data. The system uses software-based hash table implementations in RAM that consume significantly less power than hardware-based TCAM, achieving the same routing lookup functionality at lower energy cost
2Measurement precision
If multiple hash tables with different prefix lengths are created to store routing entries, then routing accuracy is improved, but the number of tables and memory usage increase
Solution Approach 1:
The patent creates a universal hash table structure that can handle multiple prefix lengths within a single table. Each hash table is designed to accommodate variable-length prefixes through bit manipulation techniques, allowing one hash table to serve multiple prefix length requirements rather than requiring separate tables for each prefix length
Solution Approach 2:
The patent merges multiple routing entries with different prefix lengths into unified hash tables. By combining entries that would traditionally require separate tables and using software-based hashing algorithms, the system reduces the total number of tables while maintaining the ability to accurately match different prefix lengths through logical operations on the stored data
3Device complexity
If prefix lengths are converted to unified length for hash table storage, then the number of hash tables is reduced, but lookup complexity increases
Solution Approach 1:
The patent changes the parameter representation of prefix lengths by converting variable-length prefixes into a unified fixed-length format suitable for hash table storage. This parameter transformation involves extending shorter prefixes with zero bits or using bitmask operations to represent variable-length prefixes in a uniform structure, enabling efficient storage and retrieval while maintaining the ability to distinguish different prefix lengths during lookup operations
Data Source
AI summary
Examples herein disclose a smaller prefix length and a greater prefix length which are identified from a routing table of various prefix lengths. The smaller prefix length is converted into the greater prefix length. The converted prefix length and the greater prefix length are combined into a hash table.