Data retrieval strategies: Retrieving data from backups overview |
 |
| 09 May 2007 | SearchStorage.com |
|
|
Backups protect the enterprise by creating a copy of important data. If regular production data is lost or destroyed through some disaster, the backup may be all that stands between recovery and bankruptcy. But the goal of backups isn't about creating a copy; it's about timely recovery. A backup administrator must be able to locate the appropriate backup media and restore the lost data within an acceptable recovery time objective (RTO). Excessive recovery time translates to lost revenue and customer relation problems. This overview outlines some of the most important issues involved in retrieving data from backups.
Setting the RTO
Recovery is a process, not an event. One of the biggest challenges with backup planning is the establishment of a realistic RTO that will vary depending on the technology in use and the amount of data to be retrieved. No two organizations have the same recovery objectives. Tape-based technologies typically have the longest RTOs -- frequently up to 24 hours or more. Disk-based technologies, such as virtual tape libraries (VTL) can have significantly shorter RTOs of just a few hours. The trick is to select a backup product that meets your recovery objectives within an acceptable budget.
Also consider the extent of retrieval objectives. That is, an organization may not need every last file back in place to resume operation. In actual practice, only a few key applications may be required. For example, rather than recovering all 2 terabytes (TB) of files and applications, the company may be able to resume operation with only a 200 GB SQL database and a handful of other applications. Any other data can always be recovered later after hours or in the background, so decide what "recovered" really means for the business.
Reducing data
Another way to accelerate data retrieval is to reduce the amount of data that must be backed up in the first place. There are several ways to reduce the overall volume of backup data. First, be more selective in the content that is backed up -- not all data is important to an enterprise. For example, .MP3 audio or .MPG video files generally do not need to be backed up. These "marginal" file types can usually be ignored unless they are specifically relevant to your business.
The other technology influencing data reduction is data deduplication, also called intelligent compression or single-instance storage. For example, a typical data center may contain 100 copies of a 1 MB document that has been shared across an organization. This would require 100 MB of storage on the tape or disk media. Data deduplication saves only one copy of a file or block to the target media, so this would effectively save one 1 MB copy of the document, issuing pointers to the other iterations -- a savings of 99 MB. Ultimately, reducing data speeds backups, and this also speeds recovery while saving money on backup media.
Media management and retention
Data retrieval is also affected by the management and storage of the media itself. Tapes and removable hard drives must be transferred to appropriate storage facilities. While media can be stored locally in a waterproof/fireproof safe, media is more frequently stored off site at a branch office or commercial storage facility, like Iron Mountain Inc. Offsite storage protects data against the threat of damage to local facilities. Media should be physically secure and stored in a temperature and humidity controlled environment. Proper storage conditions help to maintain the media's readability, especially for long-term storage. Tape also demands physical attention, such as routine drive cleaning and aging cartridge replacement. Ignoring regular maintenance and replacement may compromise the backup's readability.
Remember that offsite facilities can add cost to media storage. Media transport also introduces lag to the retrieval process. For example, it can take up to 24 hours just to retrieve a tape set from Iron Mountain before retrieval can even take place, so transportation lag must be added to the RTO.
Another factor that is often overlooked is the affect of technology changes on media. For example, moving from DLT to LTO tape drives may add greater backup capacity and performance, but previous tapes written on older drives may not be readable in the newer drives -- potentially leaving your existing backups unreadable. Readability problems may also occur when changing backup software vendors or updating backup software to a later revision. Any time that a change in backup hardware or software is planned, it's important to consider the resulting impact on existing backup media. In many cases, it may be necessary to refresh backups using the updated hardware/software.
Routine retrieval testing
Most storage organizations verify the integrity of their backups by periodically testing the recovery process. For some companies, testing is performed on an ongoing basis as lost user files are searched and retrieved. For other businesses, the testing process is more formal and comprehensive -- simulating a complete system recovery once maybe every three to six months. Regardless of the extent or frequency of the recovery exercise, periodic testing verifies the readability of your media and helps to reinforce established retrieval procedures for the IT staff.
|
|
|
|
