Replication and legacy vaulting comparison

With legacy vaulting, the most recent backups of a client are synchronized to the vault. When a new master or full backup is created, this backup is then vaulted, and the prior vaulted backup removed. With replication, previously replicated backups are retained on the target, as long as there is room on the system and the backups are within the retention and legal hold limits set on the target. See About retention control for details.

A system’s installation type governs how retention can be configured on a given system. For a description of each installation type, see Installation types and replication. Retention settings can be configured for the following installation types:

•

Local backup system to manage retention of backups run on the system.

•

Replication target to manage retention of backups replicated to the system.

•

Local backup system and replication target to manage retention of both backups run on and replicated to the system.

Retention settings cannot be configured for legacy vaults as vaulted data is not retained.

Once the initial data set has been replicated to the target, only changed data blocks are transferred. Deduplication works differently in replicating and vaulting systems.

•

In replication, the process connects to the target system and compares data in the backup to data on the target. This comparison runs on the source system. Once changed blocks are identified, they are replicated to the target. This keeps bandwidth utilization to a minimum as only changed blocks are sent through the encrypted connection.

•

In legacy vaulting, a process on the source compares backup data on the source and target to identify changes. Changes are written to a delta file that is sent to the vault, so that only changed blocks are received.

In replicating systems, backups that are encrypted on the source are encrypted on the target using the target system’s key. In-flight, the backup data is first decrypted via the transmission protocol, then before being saved on the target, is re-encrypted using the target’s key. If encryption is not configured on the target, replication of encrypted backups fails. For this reason, it is recommended that encryption be configured on the target system. Once encryption is configured, the target can receive both encrypted and non-encrypted backups from source systems for replication.

In vaulting systems, encrypted backups remain encrypted as they were on the source system. To restore encrypted data from the vault, the source system’s encryption key must be used.

Replicated data is stored on the target as a backup, equivalent to its source-side counterpart. Vaulted data is not backup data. Because of this fundamental difference, the manner in which data is restored differs significantly in replication and vault systems.

To restore from a vault, you use a disaster recovery (DR) tool, through which you first restore the system state to a newly imaged system, then restore vaulted client backups, and finally restore these backups from the new system to the client. See Disaster recovery from vault for details. To restore only a volume, directory, or file, you use the procedure Granular restore from vault.

Replication is more flexible, enabling hot/hot restore of replicated data. With replication, you can perform whole system DR, or restore replicated backups directly to a client without first restoring the entire system. For details, see Restoring replicated backups.