This article discusses data compression and deduplication on Datto devices.
By default, all data pools on Datto devices come with compression turned on. Compression saves space and improves performance. This is because time required to compress and decompress the data is less than then time it takes to read and write uncompressed data to the disk.
Datto devices use ZFS LZJB compression to compress data backed up to the device. LZJB is a lossless data compression algorithm, meaning the original data can be perfectly reconstructed from the compressed data. The compression ratio changes depending on the contents of the data itself; different algorithms get varying relative compression rates on different kinds of data.
All ZFS data, including NAS share zvols, inherits the compression settings from the zpool property; compression cannot be changed on a share-by-share basis.
Datto devices do not use ZFS deduplication. Incremental backups, however, effectively deduplicate the base image from every backup, using the inverse chain method. More information on the inverse chain can be found here.