The recent Journalspace data loss episode, is a good lesson for all of us. I don’t know what exactly went wrong, but it highlights the fact that real businesses today can’t depend upon only tape/disk based archival of business critical data.
IMO, a good data protection strategy should value following points –
In simple terms RPO (Recovery Point Objective) is – “up to what point in time the data can be recovered” and RTO (Recovery Time Objective) implies – “how much time would it take to recover data”. These goals differ from business to business. But, any real enterprise can’t depend on slow tapes for this, especially when it comes to critical customer data.
The two terms – “Backup” & “Archival” are often confused with each other. And as I see it going forward, more and more enterprises would use local disk based backup-recovery for lowering RPO & RTO. The Tape/VTL/Tape would only be used for archival of older and currently-not-being-used data for compliance or specific business reasons.
With remote IP-based replication almost becoming commodity now, it can easily be used to avoid any local disasters. But, it would be interesting to see it being integrated with local disk-based backup systems. This could reduce load on the production server and make recovery simple.
On the good side, I loved the CouchSurfing 2.0 Rised from the Ashes story covered by Luxman. The company lost the entire MySQL database almost two years back because of a faulty file-system. Founder & CEO Casey Fenton announced company’s death. Later, Fenton’s email was met with vocal opposition to the termination of the project and considerable support for its recreation.
“CouchSurfing 2.0″ was announced early in July 2006, with the intent to be operational within 10 days. The initial implementation of CouchSurfing 2.0 actually launched after only four days.