Backup and disaster recovery

Backups, Restores and Disaster Recovery
Copyright Nash Networks 2007

This much is certain: At some point, every company and computer user will lose data.

The extent and implications of that data loss will vary tremendously. You might lose an important document or a treasured photograph, or you could lose your whole business.

The bottom line: Every computer user needs to have a plan, and it needs to be tailored to your specific environment.

Backup defined

Backups are copies of documents and settings. Classically, businesses back up to tape once a day, though backup to disk is gaining ground. Tapes or disks are stored onsite or offsite. Individual users often make backups onto external drives. Backups are often only partially automated, or completely manual.

Ordinary backups only make copies of data, not operating systems or programs. "Bare metal backups" do back up everything, and allow the entire computer to be restored in the event of a major crash.

Our preference is for frequent online backup for fast-changing data - we back up hourly rather than daily. Users often lose files that took hours of work, and these can easily be restored using online backup systems.

Versions should go back at least a month to prevent losses from backing up corrupted data, or to retrieve old documents where needed. For longer-term storage, there are many possible ways to archive data.

Automation is absolutely key. Many, many, many backup strategies fail through human error.

Backups are less than useless if they won't restore

The restore, as the name implies, is the process of replacing data on a computer. "Bare metal restores" replace everything, including programs and operating systems.

Restores are the orphans of the disaster recovery process, because they are rarely tested until they are needed - and by then it's often too late.

We've seen numerous clients who diligently backed up data every day, but when the crunch came, the backup could not be restored. Software restore functions should be thoroughly researched, and restore processes must be properly tested.

Tape restores are notoriously vulnerable, because tapes may become degraded over time, and if there are multiple tapes, restores can be very slow.

Disaster recovery

A laptop is stolen, the building burns down, there's a hurricane, the central server fails. Disaster recovery is the process by which the entire computing system is restored to function. 

There are as many strategies are there are computing systems, but the basic principles are similar in any environment:

• Think through the worst-case scenario - the hurricane or the fire. What are the most critical processes and how long can you afford to be without them? What resources do you already have - offsite backups, mirror servers at other locations, offsite servers, application disks, license records?
• Now, consider the most common problems. That fire will probably never happen, but you can be certain that within the next few months you will over-write files by mistake, your computer will hang and destroy a document you've been working on all day, or your hard drive will crash. Online backup of critical data has never been easier.
• Design, implement and maintain a DR strategy. Make sure it's simple, cost-effective and that it works. Make sure it's automated and tested. Revisit the strategy regularly.