The traditional way of backing up SAP is using the Backint API. Backint APIs result in periodic full backups to either a local storage cache on the production from where the data is copied onto the backup server or to an NFS mount point where the data gets stored on the NFS share and then the data gets stored in a proprietary dedup format. But in both approaches, there is a periodic full backup and the data captured is in a backup format and not in a native block format.
Challenges associated with the Backint approach
- Performance Impact of periodic full backups: Full backups increase the I/Os and are network intensive on the production environment.
- Large RPO because of periodic full backups: Because of periodic full backups, users don’t get their desired SLAs on Recovery Point Objectives.
- Large RTO Recoveries are lengthy requiring full data copy: Since the data captured is in a backup format, users have to perform a full recovery before they can use the data.
- Large DB cloning times for test/dev: Instant DB cloning is not possible because the captured data is in a backup format.
These costs get amplified because of high compute/network/block storage costs in the cloud. What if there is an alternate approach that could give reduced recovery time for the backups?
An alternate approach
An alternate way of overcoming the challenges associated with the Backint API approach is to invoke the same savepoint API that SAP uses internally to flush data from memory to disk. Once the data is flushed to disk, in order to ensure that no new writes happen to the disk, take a snapshot of the underlying disk. Then, with a bitmap connector that keeps track of changed blocks, the product can transfer only the changed blocks from the production to the backup server. Since the data is captured in a native format, cloning data for doing test/dev becomes a lot easier with this approach.
Actifio GO is a true SaaS data protection and recovery platform for SAP using HANA, Oracle, MS SQL databases running in Cloud VMs. It delivers application consistent, incremental forever backups, instant recovery for DR, and rapid DB cloning for test/dev.
To learn more about how it works, watch a short webinar or download the white paper here.