By Scott Moore, Technical Account Manager and 20-year veteran of Ingram Micro
Having a current offsite backup of critical data in case of disaster is crucial to the survivability of any business. That being said, it is surprising how frequently this bit of protection is skipped when configuring data center solutions. “I don’t need to worry about backups; I’m using RAID and redundant power supplies”, is a phrase that should make any IT advisor worth their salt cringe. Sometimes it takes the corrosion of a faulty sprinkler head before the importance of maintaining an off-site backup begins to sink in.
As a backup mechanism for offsite storage, tape, with its long archival life, is a tried-and-true option that has been around almost as long as the digital computer industry itself. The three main points to address when selecting a tape backup solution are:
- Capacity - You’ll need to know how much data your customer needs to back up. Special consideration should also be paid to the projected growth of that data source over the next five years, or however long the customer intends to use the solution.
- Backup window - The amount of time available for the backup job to complete. For example, in an office that closes at 10 p.m. and opens at 8 a.m., the backup window would be the 10-hour period where little to no traffic is occurring on the network servers. In theory, the backup should start once business is no longer taking place, and be completed before the office starts up again the following morning.
- Backup plan - Knowing how much your customer needs to back up and how much time they have to do it in allows you to select the appropriate hardware engine for doing backups, as well as to decide how the backups should be performed.
In terms of hardware, if the end user needs to back up a handful of critical files every night that might only take an hour or two, a single LTO Ultrium drive may be the perfect solution. If there’s too much data to fit on a single tape, and the end user doesn’t have anyone to swap tapes in the middle of the night when the first one gets filled up, consider an autoloader. Autoloaders hold multiple tapes with a built-in robot that automatically swaps them in and out in front of the internal tape drive when they get full.
Afraid the backup window isn’t going to be long enough? A tape library, which holds multiple tapes like an autoloader but also houses multiple tape drive mechanisms, can cut backup time significantly by running multiple tape drives simultaneously. A backup job that takes 12 hours on a single drive will take around six hours with a two-drive library, or three hours with a four-drive library.
Another option would be to consider disk staging or a virtual tape library (VTL), sometimes referred to as a D2D (for disk-to-disk). VTLs are typically hard drive-based and can be used as a destination for backup just like tape can. In addition, since they are hard drive-based, VTLs are generally pretty easy to scale up and can be much quicker than tape when it comes to restores. Many VTLs also support some form of deduplication that can reduce backup storage needs by eliminating redundant data from the backup set.
Although VTLs offer several advantages over tape, don’t forget one of the main purposes of doing backups: the ability to maintain an offsite copy of crucial data. From most VTL devices, it is usually simple to export the data to physical tapes for offsite purposes (sometimes referred to as D2D2T for disk-to-disk-to-tape).
Finally, some consideration needs to be paid to the backup plan itself. What needs to be backed up? Is it really everything or just certain critical files? How frequently does it change and need to be backed up? For SOHO environments, a single tape might work fine and even be able to contain multiple backups. For medium, large, or enterprise customers, there might be different concerns for different segments of data.
One common backup scheme is to perform a full backup of critical data on a scheduled basis, such as once a month. Then, once a week, the changes that occurred since the previous full backup are stored as differential backups. Finally, on a daily basis, just the files that have changed since the previous backup (full or differential) are saved as incremental backups. In this way, you are backing up significantly less data than a full backup on a regular basis, resulting in a smaller backup window, but you can still retrieve any necessary files by having access to the latest full, differential, and incremental backups.
What are your customers' storage needs? Let us know in the comments below.