Secure Cache Indexing via Cryptographic Hash Functions

The patent introduces a secure hash function as an intermediary between the cache address and the cache index. Instead of directly using address bits for indexing, the address is first processed through a cryptographic hash function (e.g., SHA-256) to generate a hash value, which then serves as the cache index. This intermediary transformation obscures the relationship between addresses and cache locations, preventing attackers from inferring address information through cache timing analysis while maintaining efficient cache operations.

The patent replaces the traditional mechanical/address-based indexing system with a cryptographic hash-based indexing system. Instead of directly mapping address bits to cache indices through simple bitwise operations, the system uses cryptographic hash functions to transform addresses into indices. This substitution makes the cache geometry invisible to reverse engineering attempts, as the hash function output appears random and provides no information about the underlying cache structure or address mappings.

The patent fundamentally changes the indexing parameter from direct address bits to cryptographic hash values. By transforming the index generation process from a linear bitwise operation to a non-linear cryptographic hash function, the patent distributes cache accesses more uniformly across the cache space. This parameter change eliminates pathological conflicts that occur in traditional indexing where specific address patterns systematically collide into the same cache indices, thereby improving cache operation reliability and performance consistency.