Data deduplication vendors constantly tout two big data deduplication selling points: higher reliability and faster restores (compared to tape). No doubt about it, these are important. And, we’ve seen some exceptional results with data deduplication practice in production environments of large and mid-size enterprises.
However, vendors are missing the biggest bang for the buck especially in large enterprises—improving disaster recovery with remote vaulting. Many large enterprises have always wanted to remote vault more of their data offsite. And, many are replicating primary data, and/or creating snapshots or performing continual data protection (CDP) for data on primary storage. However, moving 10s or 100s of TBs of nightly backups that are currently going to tape requires large communication pipes to electronically vault the data to a remote site, and it is often too costly to justify. In addition, the requirements for getting data offsite are normally under the same stringent SLAs as recovery time objective (RTO) and recovery point objective (RPO). In order to accomplish the offsite requirement, deduplication appliances can play a part in getting backup data offsite in an automated process.
Dedupe solutions for backup operations can greatly reduce communications costs while enhancing manual and labor intensive processes. Deduplication technology reduces the communication line needs by factors of 20-50+:1. Only metadata and new blocks need to be replicated daily. In addition, backup software has evolved to support these needs as well by automating the replication process, recognizing the original as well as the replicated copy, archiving long retention data to tape, and cataloging the complete processes.
Disaster recovery is enhanced with data deduplication. Organizations can carry out remote vaulting via automated processes, increase reliability, speed up restores, and reduce recovery point and time objectives. As organizations assess the value of data deduplication to their organization, disaster recovery should be a part of the equation—a big part.