Replication Of Incremental Backups - HP 12000 Design Manual

backup tape on the source device (not the tape currently replicating) runs in parallel with the
replication.
5. Once the backup has successfully replicated, the backup is reassembled using the transferred
deltas and the older existing duplicate components on the target device.
6. Space reclamation of the previous backups (not the tape just replicated) on the target device
begins. The target runs its own space reclamation to eliminate the duplicate data on the older
backups on the target.
Figure 45 The Replication Sequence
1. A backup is performed (and its metadata is stored in the database).
2. The new backup is compared to the previous versions and the unique data in the new backup
is identified.
3. The source device performs space reclamation on the previous backups to eliminate the
duplicate data, and concurrently the replication data (containing just the unique data and
metadata) is transferred to the target device.
4. The replication data is reassembled into an intact backup combining the new unique data
with the duplicate data already replicated from previous backups.
5. The target device performs space reclamation on the previous backups to eliminate the duplicate
data.
HP VLS backup systems with HP Accelerated deduplication allow only unique pieces of data to
replicate across the WAN link. This efficiency in understanding precisely which data needs to
replicate may result in bandwidth savings in excess of 90% compared to transmitting the full
contents of a cartridge at the source site. In addition to the unique data transmitted, there is some
overhead of data instructions that also needs to pass across the replication link. This is known as
control data, and is estimated to be about 1% of the full backup size. The control data is optionally
compressed before it is transmitted. There are also some variances in the amount of data replicated
from database backups and file system backups because of the different deduplication algorithms
used and differences with incremental backups.

Replication of Incremental Backups

Most database incremental backups consist of block level changes in the database, while most
file level incremental backups consist of changes at a file level. Depending on your recovery point
objective (RPO) requirements at the disaster recovery site, you may choose whether or not to
replicate incrementals. With the current capacity-optimized comparison algorithms in the HP VLS,
the amount of data replicated for file system incrementals is almost the entire size of the incremental
How it Works 99