Compression; Space Reclamation; Remote Replication - Quantum DXi4700 User Manual

Chapter 1: System Description

Space Reclamation

Quantum's deduplication technology uses a sub-file, variable-length approach to identify redundant blocks
in a data stream—blocks that have appeared before in the same dataset or in datasets processed at an
earlier time. When a block appears that has already been stored, the DXi system inserts a reference pointer
to the earlier instance of the data segment instead of storing another copy. The result is a dramatic reduction
in the storage capacity needed to store the data set, and a similar reduction in the bandwidth needed to
replicate deduplicated data sets over a network.

Compression

The DXi4700 systems use compression technology after duplicate blocks have been identified and replaced
as part of the deduplication process. With compression, unique data that has been through the data
deduplication process can be compressed at a typical ratio of approximately 2:1. This enables you to
maximize the storage capacity of your system.

Space Reclamation

The space reclamation process performs multiple functions on the DXi4700.
When data is deduplicated it is stored in a block pool—a pool of all unique data blocks that were captured
during the data deduplication cycle. When subsequent backup jobs occur, the data deduplication engine
searches for new data entering the DXi and uses a variable length compression algorithm to compare new
data to existing data in the block pool. Unique blocks are added to the block pool and known blocks are
indexed.
The space reclamation function searches the blockpool for data blocks that are not referenced by any
pointers (that is, the files associated with the block have been expired and removed). Once such a data block
is identified, the block is removed to make the space reusable.
For correct system operation, space reclamation must be run at regular intervals (at least once a week).
Quantum recommends creating a schedule to automatically run space reclamation (see ).
It may be beneficial to schedule space reclamation for a time when other operations are not normally being
carried out. Therefore it is important to know when to schedule the space reclamation process. As best
practice it is recommended that this process commences at least two hours after your backup job has
completed on a daily basis. It is far more efficient to process a day's worth of new data than a week's worth.

Remote Replication

Today most backup occurs on isolated devices, making it difficult to deploy disk backup when disaster
recovery protection is required. DXi-Series solutions use data deduplication and replication to decrease by
up to 50 times the bandwidth required to move backup data over networks and between sites. This dramatic
Quantum DXi4700 User's Guide 20