English
 | 
Deutsch

Data security  

  Search
   Home  > >  IT-Services  > >  Data security

Backups without functioning restores are useless

Many clients proudly show us their forbiddingly expensive tape silos and at the same time forget that their backups are only as good as the restores. But restores are almost never tested – much less a disaster recovery.

It took one of our customers more than two weeks to restore an exchange server critical to the business. Only thanks to the upcoming Christmas break was he able to prevent reduced hours for thousands of his employees.

Backup Restore & Disaster Recovery

Even today, traditional backups using tape drives cannot yet be completely replaced for larger installations. Besides cost reasons, Fiberworks recommends a combination of different technologies to achieve optimum data security and cost.

The two most popular variants are weekly full backups with daily incremental or differential backups. For these two procedures, the data is backed up fully once a week. The daily changes in data are recorded and stored on additional tapes. In case of a restore, the full backup first has to be played back , followed by all incremental backups.

This approach is not deterministic, i.e., it cannot be predetermined when this type of restore can be completed successfully. Complex environments often operate multiple tape drives in parallel to achieve a greater throughput and to realize improved backup times. Backup jobs are run throughout the week to optimally utilize the hardware. In case of a restore, the restore can only be started if sufficient tape drives are available. In addition, all incremental backups my be distributed across hundreds of tape drives. Each tape has to be loaded first, then wound back to the start of storage while no tape or write errors must exist as else the whole process would have to be restarted using a replacement tape.

Restores in complex environments may take DAYS to complete.

This type of approach is therefore totally inadequate to protect your primary business processes. Fiberworks recommends the definition of SLAs based on your business processes. The SLAs will define the technology to be used, the type of restores and thus the backups - and not the other way around. Moreover, the scenario described here deals only with the application data and not with the operation system.

To be able to start a  restore, the following steps will have to be performed in case of total failure (disaster recovery):

1. Procure new hardware
2. Host bus adapter installed and configured
3. Install operating system
4. Configure network and SAN
5. Install and configure backup program


Many clients proudly show us their forbiddingly expensive tape silos and at the same time forget that their backups are only as good as the restores. But restores are almost never tested – much less a disaster recovery.

It took one of our customers more than two weeks to restore an exchange server critical to the business. Only thanks to the upcoming Christmas break was he able to prevent reduced hours for thousands of his employees.

For data that is critical to your business,  Fiberworks recommends the backup to disk procedure.

Our experts would be happy to advice you. Based on your business process defined via SLAs, TIER level and possible virtualization potential, we will develop a backup, restore and disaster recovery strategy tailor-made to your company's specific needs.

The classification of the business processes as well as the creation and verification of SLAs does not only guarantee smooth and secure operations but also helps significantly cut costs. An optimum level can be achieved, if a TIER level and backup processes are considered equally.

Interested? – Contact us

Copyright 2002-2008 by Fiberworks
Friday, November 21, 2008  |  Contact  |  Imprint  |  Login