Redvers Compression Algorithm - Case Study
On April 15th 2009 Redvers Consulting conducted two COBOL application simulations at the IBM Innovation Centre in Hursley Park, Winchester, UK. The simulations ran on a 2-way Power 5+, 16GB System i server running IBM i 6.1.
The first simulation was to prove the Redvers Compression Algorithm could be used to reduce the amount of memory required to hold one or more instances of a string of 1,214 bytes. The software also had to restore the compressed string back to its original form and confirm an exact match.
The results showed a memory saving of 751 bytes - a compression ratio of 62%.
The second scenario simulated the batch compression and decompression of a client master file consisting of one thousand 1,212 byte records, each with a 16 byte key.
To allow for easy access to the compressed data, this simulation did not compress the record keys. Only the data portion of each record was compressed and a two byte length was added after each key, providing a length for the compressed area.
Finally, the compressed master file was restored to its expanded state and checked against the original file.
These results proved that a 1.2MB master file can be converted to a 0.5MB file with equal accessibility and data content - a compression ratio of 60%.