Backup to disk is fundamentally changing corporate perceptions about backup and recovery. Using disk as a primary backup target has solved long-standing corporate backup problems including successfully completing backups within designated backup windows and expediting recoveries while deduplication is resolving the cost and capacity issues associated with storing backup data on disk. But before companies breathe a collective sigh of relief and think that disk has officially solved their backup problems, they need to think again. The immediate crisis may be over but longer term problems still remain.
Storing backup data on disk creates its own set of longer term problems that can come back to haunt companies in ways that are only now becoming understood. Specific problems that companies may need to manage include:
- The length of time to retain backup data on disk after the primary production data is deleted. Companies and/or employees from time to time purge production data (emails, files, database records, etc.) due to space constraints, regulatory requirements, as part of internal practices for data management practices or even by accident. But if and when the data is deleted from the production data store, how long should the backup copy remain on disk?
- The increased energy and cooling costs of using disk to store backup data. Deduplicated or not, deduplicated backup data stores will likely keep growing and storing data on disk requires power and cooling for these storage systems. With power costs everywhere on the upswing and some companies even finding that more power is not an option, companies may find they may need to introduce and migrate data to MAID (Massive Array of Idle Disks) storage systems into their environment in order to keep data on disk while still controlling power costs.
- Prioritizing what backup data is kept on what type of disk. Minimizing power costs is only one reason to store data on different types of disk. Other concerns are speed of recovery and controlling storage costs. SATA disk storage systems are great for maximizing capacity while minimizing costs but they don’t always perform as well as FC or SAS disk drives. Conversely, when companies need to restore data in a hurry, such as during a disaster, keeping some backup data for certain applications on high performance FC or SAS drives may make sense for these applications to expedite their recoveries.
- Searching and retaining backup data to satisfy e-discoveries and legal holds. Recent changes to the Federal Rules of Civil Procedure (FRCP) apply to every company regardless of its size and backup data stores are just as subject to e-discoveries and legal holds as data in production. How easy and cost-effective it is to access, search, retrieve and retain backup data on disk are already issues and bound to become more important over time.
- Migrating backup data when technology refreshes of disk storage systems occur. No disk-based storage system lasts forever and most companies will find it advantagous to migrate backup data regularly (every 3 -5 years) to new storage systems to ensure their storage systems are under warranty as well as to take advantage of advances of higher capacity disk drives and the lower prices that accompany them. How easy it is to manage the migration of backup data to new disk storage systems and what data they should migrate are factors that companies need to consider when implementing any disk-based storage system.
Obviously more issues than this exist as part of using disk to store backup data for the long term (over 30 days), but these points illustrate the importance of selecting software that does more than just stores backup data and deduplicates it. These issues should prompt companies to look at more deeply at what software they are going to use in conjunction with disk in order to mitigate these longer term issues associated with backing up data to disk. In a series of upcoming blog entries, I’ll take a deeper look at the new Backup Lifecycle Management (BLM) Archiver feature of Asigra Televaulting 8.0 and how it specifically addresses issues associated with the management and placement of backup data stored to disk short and long term.