MANAGE AND SHARE DATA

BACKING UP DATA

Making your own back-ups is an essential element of data management. Regular back-ups protect against a raft of potential mishaps, the most obvious of which is data loss due to hardware failure. But there are others. Back-ups protect against accidental and malicious data loss due to:

• software faults
• media faults
• virus infection
• malicious hackers
• power failure
• hardware faults
• fire, flood or other severe catastrophe

The main principle of backing up data is that copies of critical files are made which can be used to restore originals if the data suffer from one of the above forms of data loss. Regular back-ups can also help to prevent data loss, especially:

• accidental change to data
• accidental deletion of data

The precise back-up procedure you should adopt depends on local circumstances, the perceived value of the data and the levels of risk which you consider to be appropriate for your circumstances. The type of media you need to most effectively make back-ups depends on the quantities and types of data which are being backed up.

For many researchers, carrying out an informal risk analysis provides an indication of back-up needs. If you have read this far down the page you probably already feel that the costs of replacing critical data files will be greater than the costs of implementing some back-up procedures. There are few hard and fast rules for backing up, but the following, significant issues should be addressed.

Content of back-up

What will you need to restore in the event of data loss? If your institution can restore your system then you may wish to take responsibility only for your data files. If it cannot, you may wish to take full responsibility for your own 'system' back-ups.

Frequency of back-up

To reduce risk to the absolute minimum, back-ups should be made after every change to data but in practice this is near to impossible. Finding a reasonable balance which takes into account the levels of risk which you face is a personal decision, but backing up frequently used and critical data files daily minimises the effects of any data loss event. This is best achieved using an automated back-up process. Microsoft SyncToy is an easy-to-use method of synchronising files in different locations.

Incremental or differential back-ups

An 'incremental back-up' is based on first making a copy of all relevant files (often the complete contents of a personal computer) and then taking incremental back-ups of the files which have altered since the last back-up. 'Differential back-ups' differ as, while a complete back-up is made first, all subsequent back-ups deal with the files changed or created since the first full back-up and not just since the last partial back-up. In terms of resources, incremental back-ups can be made on removable media (i.e. CDs, DVDs etc.) while differential back-ups are more suited to 'fixed' media like hard drives. Whichever method is used, it is best not to overwrite old back-ups with new.

Media choice

The choice of media depends on the quantity of files and the preferred method of backing up. If you are backing up a hundred small data files on a daily basis, just copying them to a recordable CD probably suffices. If, on the other hand, you are making differential back-ups of very large quantities of data using a networked hard drive, a removable hard drive or even magnetic tape is probably more convenient. Recordable CDs and DVDs are now so cheap that replacing media regularly should no longer be a cost issue.

Location of back-up files

Depending on the form of back-up and the risks associated with data loss, it is generally most convenient to have 'online' back-up files. However, if the online back-up is in close proximity to the 'at risk' files, then a similar chance of data loss probably prevails. For critical research data which are not available elsewhere, we would recommend that offline storage (i.e. files held on removable media or transportable hard drives which can be physically removed to another location for safe-keeping) is a minimum requirement. Most manufacturers provide recommendations for the best storage conditions of physical media.

Institutional policy

Most universities have a back-up policy for networked files. If your data are held on a university network space, then this may be backed up regularly by the university. If all your data files are stored here, then it is worth checking with the university about what its policy is, and if you feel that there is any risk, you should maintain independent back-up of critical files.

Validation of back-up copies

It is important to verify your back-up files regularly by fully restoring them to another location and comparing them with the original.

Organisation

If you are making your own back-ups on removable media, make sure they are well labelled and well organised. Managing your back-ups is just as important as making them. Without some management, achieving the ultimate aim of restoring lost data may prove difficult.