Tapes have been around for decades, especially for storing high volumes of data. Disk based backups have also been around for years; however, they have been fairly expensive per terabyte (TB) of data. Today you can find virtual tape library (VTL) solutions for roughly double the cost of that of a physical tape library, per TB. Disk based backup technology pricing has progressed to a point where it can seriously compete with traditional tape solutions. The price gap may seem quite large for some, but when one considers the operational overhead associated with maintaining tape backups, the gap becomes negligible.
Tape based backups are notorious for their high operational overhead attributed to high failure rate. Tapes constantly require refreshing, tape drives need cleaning and tape libraries need to be restocked with tapes. Not to mention the environmental considerations for storing tape media. Failure to store tapes at the appropriate temperature and humidity will eventually lead to poor data retention and a higher backup failure rate. It is my belief that tape technology still plays an important role in data protection; however one may consider not using it as a primary backup storage medium. The amount of mechanical stress placed on tape drives and media is too great to promote reliability of the backed up data and diminishes the probability that the data will be recoverable. Tapes are sequential and ideal for scenarios where a constant stream of data is provided. Failure to provide that stream will cause the tape drive (and tape) to wear down and lead to premature failure. Furthermore, tapes have derived a poor track record not because it is an inferior technology but usually because of improper operator maintenance and non-optimal environment configuration. Tapes are very useful, and most times a necessity, for archiving large volumes of data to maintain governmental compliance.
VTLs aren't necessarily for everyone and may not be exclusive for organizations which require sending copies of their backups to offsite storage facilities.
What are some advantages of a VTL?
- VTLs provide a random access, instant mount/position storage medium, which will decrease your backup windows. This is extremely important if you have many slow clients. Backing up to disk will no doubt also increase your backup and restore reliability; hence you will be spending less time managing your environment.
- Expect to spend less time having to clean tape drives, restocking tapes and “up’ing” tape drives. Technological advantages including data deduplication.
- Many VTL solutions now provide efficient in-line data deduplication (also known as “de-dupe). VTLs which handle in-line “de-dup” (data deduplication) will reduce the necessity to store redundant blocks of data and may increase your overall ROI for migrating to a VTL. If your incremental backups are consistently small (up to 10% of your entire backup set) you may see a huge benefit from in-line de-dupe.
- VTLs can also be replicated across remote sites, eliminating the need and risk associated with transport of offsite data.
However, VTLs aren’t without their shortcomings. Capacity will not be as easily scaled as with physical tapes. VTLs cannot be sent offsite for long-term storage as with tapes. Thus if you still require tape off sites, you will most likely be looking at a hybrid solution.
What are common considerations when migrating to a VTL?
- Proper capacity sizing is critical, because adding more storage down the road will not be as simple as purchasing new tapes.
- Tally up all the data you currently have stored on tape, which has not expired. You will more than likely find that your storing more data than you will ever need or are required. Based on your budget, you may need to reduce the total number of older backup sets you store on the VTL and store them on tape. This can be accomplished using tape duplication (VTL to Tape), before the data expires on the VTL, and updating the retention of the duplicate image. In any case, the VTL will still be used as your “staging” environment thus providing a constant stream of data to the tapes and increasing your success rate. You may also consider directly connecting high volume servers to your VTL and avoid network bottlenecks.
Other considerations are feature based like “OST” and “de-dup”. Choosing a VTL which supports OST (Open Storage) will reduce some operational overhead associated with managing expired data. VTL solutions are not immune to vendor lock-in. Careful consideration should be taken to determine if you want an appliance or a storage agnostic environment. Most VTLs are fiber attached and will require a fiber infrastructure to support them. Depending on the VTL, you may have to upgrade your power infrastructure and incorporate seismic restraints to support your VTL. Finally after you’ll want to let your “old” tape data expire and not migrate it to the VTL. So plan on maintaining support for your ATL hardware and tape storage well after you migrate to a VTL.