Backup infrastructure forms a critical component of the information technology (IT) setup of any organization. To protect their most critical asset — data, organizations today are deploying backup software as well as new technologies like deduplication-enabled backup devices. However, in order to get the most from the backup infrastructure, the
- The most important backup policy is categorizing data according to the criticality,
recovery time objectives (RTOs) and recovery point objectives (RPOs).
- If you are using a tape library as a backup device, then your backup
policy can use the following pointers:
- Data that is most critical and has the least RTO, should be fully backed up daily so that at the time of recovery, you don't have to look for incremental tapes and combine the incremental data with the recent full backup.
- Weekly, full and daily differential levels should be used for data with medium criticality. Such a backup policy ensures that recent full and differential backup tapes are only required at the time of recovery.
- If the RTO is negotiable, data should be fully backed up once a week, and backed up daily with an incremental level.
- Media pools should be defined in the backup software according to the data type. For instance, a separate media pool should be used for critical applications like SAP and Oracle. If separate media pools are not defined for critical applications, backup from different applications will be backed up on the same media. While recovering data, the RTO will increase as data from various applications are scattered across the tape.
- Separate media pools should be assigned for daily, weekly, monthly, and yearly backups. It the daily and yearly backups are taken on the same tape, it won’t expire, and can’t be reused until every yearly backup image has expired.
- A dedicated pool should be used for the catalog backups so that data can be recovered as soon as possible, in case of disaster recovery.
- In case of SQL or Oracle backups, transaction logs should be backed up hourly for minimal RPO.
- Use staging in case of disk-based backups, along with tape libraries. This entails backing up
the data to a disk-based device up for a particular duration and then staging it to tapes for
longer retention. This backup
policy will help in recovering the recent data from disk; therefore, the recovery time will be
far less as compared to recovery from tapes.
- The incremental data can be backed up to disk as part of your backup policy. The backup on disk
is faster, and can be performed along with full backups to tape. This part of the backup policy
will help you achieve narrow backup windows during the
weekdays. It also ensures that backups do not run into production time.
- The tapes can be cloned from virtual to physical tape libraries so that the data can be recovered faster, if it has not passed the retention on the virtual tape library.
If the backup policies are not designed carefully, the backup infrastructure may not yield the desired results and the organization may end up spending extra money on it. For instance, if an organization chooses to back up full data each day, it may end up spending a lot of money on tapes and backing up the same data, irrespective of its importance. Hence, categorizing data as part of the backup policy helps in fully backing up critical data each day so that the recovery time is minimized and incrementally backing up data for which the recovery time is negotiable. This also helps save on the number of tapes required and the involved costs.
About the author: Anuj Sharma is an EMC Certified and NetApp accredited professional. Sharma has experience in handling implementation projects related to SAN, NAS and BURA. He also has to his credit several research papers published globally on SAN and BURA technologies.
This was first published in December 2010