The most common value proposition that I have seen virtual tape library (VTL) vendors present has been seamless integration with existing data backup/restore infrastructure and processes. However, over the years, I've discovered that the theory has not effectively translated into reality: VTLs have not been as seamless an integration exercise as VTL vendors would have you believe.
In spite of this, VTL adoption was on the uptick over the last few years, largely because data backup/restore software at the time did not have robust native backup-to-disk capabilities; therefore, the effort to implement a VTL was tolerated for its backup-to-disk benefits.
However, backup software products today -- such as those from CommVault, EMC Corp., with its NetWorker, IBM Corp. Tivoli Storage Manager (TSM) and Symantec Corp. Veritas NetBackup -- can perform backup to disk in a more sophisticated and seamless way. In addition, with data deduplication becoming the new paradigm for backup to disk -- because of its compelling cost per gigabyte that is comparable to tape -- VTLs that don't incorporate data deduplication are becoming even less popular
Right now, the most compelling advantage of VTLs over LAN-based backup-to-disk solutions is their high-performance capabilities. Virtual tape emulation is the main backup-to-disk software interface within backup storage appliances that use high-performance Fibre Channel as the connectivity. However, this performance capability comes at a significantly high cost per gigabyte. Thus, in order to compete, VTL vendors are introducing hybrid products with VTL front-end and dedupe back-end storage. These are good solutions that provide the best of both worlds; but it's important to note that, in my findings, storage professionals who select these products are not typically making their selection based on the merits of the product's VTL interface, but on whether the solution addresses their performance and data deduplication requirements.
So what's the future? There's speculation that in a few years, when target-based network-attached storage (NAS) data deduplication products begin to perform as well as Fibre Channel VTLs, VTLs will be phased out in favor of NAS-based 10Gb Ethernet deduplication backup-to-disk solutions because of their significant ease of use, cost effectiveness and integration with native enhanced backup software features. In the future, storage professionals might end up asking themselves: "Why should I bother acquiring a backup-to-disk product that has a VTL software layer that I have to manage, when a straight backup-to-disk alternative exists that is just as fast?"
However, data deduplication as the conqueror is just one possibility out of many, even if it is a likely outcome. After all, this is the IT industry, where the uncertainty principle reigns supreme and just when you think you've got your finger on it, everything changes (solid-state drives anyone?). So, my advice today for storage professionals is to focus your efforts on aligning your requirements and recovery objectives with the overall capabilities offered by the backup-to-disk solutions you're considering. Don't get hung up on sales pitches focusing too much on specific product features presented in absolute terms (e.g., be skeptical of statements like "VTL is the best backup-to-disk approach" or "there is no point doing backup to disk without dedupe").
Having said that, what I have found is that most storage professionals that have successfully implemented a backup-to-disk solution (VTL appliance or dedupe appliance), had put the presence of a virtual tape interface as a secondary criteria to the solution's overall performance and capacity optimization capabilities.
About the author: Ashley D'Costa architects and designs advanced computer solutions and has technical experience with a broad spectrum of IT infrastructures.