An approximation to the greedy algorithm for differential compression of very large files

Abstract

A new diffrential compression algorithm that combines the hash value and suffix array techniques of Manber and Myers, was discussed. Differential compression algorithms did not find the longest matches for every offset of the version file. The performance of Greedy algorithm was based on the block hash table, the quick index array and the pointer array. Computing the suffix arrray of block hash reduced the storage and computing requirements by a factor equal to the block size used. It was found that the differential compression method was used in compression applications, web crawling as well as computational biology.