What you will learn: Data retention is an important consideration with a VTL, as the space available for backups is, unlike in a tape library, a finite resource. Therefore, developing a data retention strategy is critical for VTL users.
A virtual tape library (VTL) emulates a tape library within a disk storage system. The disk-based approach offers vastly improved speed and reliability, while remaining fully compatible with existing backup software and internal backup processes. However, whereas tape library storage is virtually infinite, disk storage is a finite resource. Space limitations will impact the retention capabilities of any VTL system.
VTL storage space is limited
The issue is storage space. Tape libraries use removable media -- the tape cartridges. If a tape library fills its tapes, it's a simple matter to insert new tapes or use an autoloader to draw spare tape cartridges from its library. The filled tapes are typically removed and sent off site anyway. In a sense, tape library storage is virtually infinite, so the retention period for tape cartridges can also be infinite.
VTL platforms do not have this advantage. Although a VTL may be upgraded with additional disk storage or retrofitted with larger capacity disk drives, the total disk storage within a VTL is finite. Consequently, as the VTL fills with new backup jobs, the old backup jobs already on the VTL will need to be deleted or moved to another storage system.
VTL storage is time
At the simplest level, the amount of time that you can store a backup job on a VTL depends on the size of the backup jobs, the frequency of the backup jobs and the capacity of the VTL.
For example, if a typical set of weekly backup jobs requires 100 GB, and the VTL offers 500 GB of storage capacity, you'll fill the VTL in five weeks. By week six, you'll either need to remove the backup job from week one or add more disk space. This is an important consideration, but is often overlooked by VTL buyers.
The solution is often to offload the older backup jobs to another storage platform (usually the tape library that your VTL replaced). Most VTLs provide support for legacy physical tape libraries and can automatically move older backup jobs in an offline fashion so that network performance is not impacted.
Data reduction extends VTL retention
Another means of extending VTL retention time is to reduce the volume of data that must be stored. Traditional compression is a popular method of doing this. While the amount of compression is relatively low (usually about 2 to 1), this almost doubles the effective storage space on disk.
Delta differencing is another technique that can extend VTL disk space. Rather than executing full backups, delta differencing only saves changed data. For example, a 100 GB backup may be run initially, but if only 10 GB of data changes during the week, only that 10 GB difference is then backed up to the VTL.
Data deduplication features are now appearing in backup software products, and this has had an enormous impact on VTL storage and retention. Average compression ratios are about 20 to 1 and can be as high as 50 to 1. But if you consider the 500 GB VTL in the example above, a 20-to-1 deduplication will turn a 100 GB backup into a 5 GB backup. Instead of fitting five 100 GB backups, the VTL could store up to 100 backup jobs, and the potential retention would jump from five weeks to almost two years.
Stephen J. Bigelow is the features writer for SearchStorage.com
