Compression reduces the amount of physical storage required for a volume by combining data blocks in compression groups, each of which is stored as a single block. Reads of compressed data are faster than in traditional compression methods because ONTAP decompresses only the compression groups that contain the requested data, not an entire file or LUN.

You can perform inline or postprocess compression, separately or in combination:

Inline data compaction

Small files or I/O padded with zeros are stored in a 4 KB block whether or not they require 4 KB of physical storage. Inline data compaction combines data chunks that would ordinarily consume multiple 4 KB blocks into a single 4 KB block on disk. Compaction takes place while data is still in memory, so it is best suited to faster controllers.