In an age of de-duping VTLs we’re stuck with tape-based software
Why does backup software still have its obsession with tape? The backup applications are riven through with the assumption that the ultimate end-target is tape, however much they nod to disk-to-disk (D2D) or virtual tape libraries (VTL).
With organisations that have abandoned tape or used it for their last-tier backup, the backup applications make a dreadful hash of managing D2D and VTL targets, especially with de-duplication. A backup application cannot write a de-duped file to tape and has to expand it before it can do so. But it is a nonsense for the same approach to be taken by a backup application when directing a de-duped storage device to move its data to a remote de-duped storage device. That negates all the bandwidth reduction benefits of de-dupe.
Backup applications have to drop the one (tape) size fits all approach and realise they are in the data protection business and that means dealing with disk-based protection targets and processes natively and not as surrogate tape devices.
As Michael Hammer might have said, customers don’t want backp software; they want data protected.
In a world of disk storage devices the tape reel barcode should not be relevant.
Disk storage arrays are getting more intelligent; they can make snapshots of their contents, they can clone them, they can replicate them. Wide Area File Services devices can backup data from branch and remote offices to data centres. Data centres can replicate files hundreds of miles to remote data centres for disaster recovery.
If we were starting out now to protect data we would not write backup software. We would write software that recognises the nature of intermediate and ultimate backup targets and manages data movement to and from them much, much better than it does at present.
We live in a data storage nuclear age but protect that data with Iron Age software. The backup software vendors have to radically develop their products or face confinement to a backup-to-tape niche as other applications take on the role of directing data protection processes and data flows.
We need, I think, profoundly new data moving software to transfer data from servers’ storage arrays to target data protection devices and profoundly new software to provide a reporting and management environment on that protected data. It may be the case that the Aptare, Bocada and EMC/WysDM data protection management (DPM) applications of this world start taking on so data protection initiation responsibilities as well.