Exact Match Hash Lookup Databases Network Switch Devices

The patent divides the hash table into multiple segments or regions, each managed independently. By segmenting the table, the collision probability within each segment is reduced, allowing more keys to be successfully inserted while maintaining exact match lookup capability. This segmentation approach resolves the contradiction by enabling larger table capacity without sacrificing reliability.

The patent introduces additional dimensions to the hash table structure, such as multiple hash functions or multi-level indexing. This dimensional expansion allows the system to accommodate more keys by distributing them across multiple dimensions, reducing collisions in any single dimension while preserving exact match lookup performance through coordinated access across dimensions.

The patent pre-computes and stores hash values for multiple hash functions during the table population phase. This preliminary action eliminates the need to compute multiple hash values during lookup, as the pre-computed values are already stored and can be directly used for collision resolution. This resolves the contradiction by improving memory utilization through multi-hash functions while avoiding the complexity penalty during the critical lookup operation.

The patent creates and maintains multiple copies of hash values corresponding to different hash functions within the table structure. These copied hash values are stored alongside the key-data pairs, enabling efficient collision resolution without re-computation. This copying strategy improves memory utilization by enabling better key distribution while keeping the lookup process simple by directly comparing pre-stored hash copies.