Backups

W e manage everything. hardware, tape labelling, calendars, verification reports and disaster recovery testing. We even phone you if you forget to put the right tape in.

We provide:

• Tape hardware;
• Standby hardware in case your fails;
• Calendar management;
• Tape labelling and media tracking;
• Monitoring tape usage and advising when media will reach end-of-life;
• Daily telephone calls in case you forget to change tapes;
• Regular Disaster Recovery testing;
• Encrypted backups (so you can safely take tapes off-site and not worry about data falling into the wrong hands); and
• Fully managed restores—just phone our support team with your request.

How our backup service differs

Research shows many basic I.T. companies will install and setup a backup system, then leave it run and only resond if someone calls to report a fault. Sometimes backup checks are included in yearly or bi-yearly healthchecks.

This differs to the 8Networks' backup service where we proactively monitor your backups on a daily basis, and guarantee that agreed data areas will be backed up according to an agreed schedule. In the event of a failure, we will personally notify you so that alternative action can be taken. The end result is that you, a Director or business owner responsible for your organisation's data, will have 100% faith that your backups will be working as expected, and you will be able to recover critical data in the event of hardware failure, OS or application error, flood, fire, theft, virus attack, accidental user overwrites, etc.

A common misconception

Backups are rarely needed to recover from a hardware failure. Of course this is a possibility, but in 9/10 cases we need to restore a single file due to user error, e.g. accidental deletion. Sometimes this is a file that hasn't been used for a number of months, for example a spreadsheet used yearly when preparing annual accounts. This is why it is important to take historical backups as well as daily backups. Many small businesses without a proper backup strategy backup weekly to a USB drive. This will help in the event of an overnight catastrophic failure, but this will not help recover a file last seen 3 months ago.

What about RAID?

RAID is not a backup system. RAID is designed to reduce the impact of a hard drive failing by providing a spare; this lets you keep working while a technician can get to you to replace the failed drive. It also means a complete recovery of your system is not required in the event of a drive failure. RAID provides no protection against deletions or overwrites and will not allow a file to be recovered due to human overwrites, accidental deletions, system errors or Viruses.

A true story

We were called once by a small business who claimed their server had put all their data back to a year ago. What was even stranger was the client was only doing weekly backups, and didn't have any tapes from a year ago, so it couldn't have been an accidental restore from the wrong tape.

A quick site visit determined it was true, the latest files were indeed a year old. After a bit of investigating, we determined the following sequence of events.

1. Their I.T. support company installed a new server with RAID and a weekly backup system
2. RAID failed within a week, but went completely undetected (RAID monitoring software not installed, and no process was setup to check such an occurence). For the past year, the company was using the server on a single disc.
3. Company does weekly backups as per instruction, but over time new software is installed that uses new data areas, and these new areas are not backed up. No processes are in place to ensure data can only be saved to areas that are backed up.
4. Over time the backup failure itself fails, and the company was putting in a tape each week but no backup was occuring. No verification checks or disaster recovery testing was performed.
5. One of the hard drives in the RAID pair fails, and this failure causes the already failed RAID system to use the other disc (last used a year ago). As a result, after a server reboot the company sees their system exactly as it was a year ago.

Why this wouldn't happen with 8Networks support

Three reasons:
1. RAID checks are part of our pro-active monitoring, so we are alerted instantly when RAID is not operating as expected.
2. We agree formal storage areas with our clients, then configure the network to prevent data being accidentally stored on areas that are not backed up.
3. We include a full Disaster Recovery test twice a year, that is a full rebuild on one of our standby servers that you sign-off as being correct.
4. We provide personal daily monitoring of your backups (and maintain a log of verification reports for compliance with various regulatory bodies).

Restore times

An important factor to consider is how quickly a file can be recovered. Our final backup is often to encrypted tape (so it can be taken off-site), but tape takes a long time to read, and often the entire tape needs to be read before the required file can be put back where the user can access it. We use a disk staging area before streaming to tape, which means at least the previous night's backup (and typically 2 or 3 other backups) can be instantly accessed. This means we can provide file recovery within you standard Service Level Agreemnent response times, and very often we can provide immediate restores while you are on the phone to us.

Off-site backups - preventing against accidental loss or theft

Almost all of the businesses we speak to are aware of the need for off-site backups. However, when we ask what would happen if that tape were to be lost or stolen, well, you can guess the response we often get.

There is no point investing in a secure network if you are taking the entire copy of your server, off-site, without encruption. This is why 8Networks encrypt all your data before it is streamed to backup tape.