Activity-based reduction reduces the performance cost incurred by decompressing frequently accessed data. This function allows up to 20 percent of the busiest data to be stored uncompressed on the system. This ability benefits the system by minimizing performance latency that results from constantly decompressing frequently accessed data. To determine which data is the busiest, the system uses ML algorithms that process I/O statistics. Activity-based reduction maintains a balanced, optimal environment for both data reduction savings and performance.
