Backing up data

Data… A few megabytes of spreadsheets or a terabyte of design assets. If your system crashes, you’ll be lost without it… So back it up.

To piece together a backup strategy, you need to understand the common approaches to backing up, and the strengths and weaknesses of available backup mediums.

Backup approaches

Full backups

A full backup creates a complete copy of an entire system, or a complete copy of selected files and folders. The result: an exact clone of selected data.

During every new full backup the entire data set is copied again, regardless of what data has been added or changed since the previous backup.

A new full backup either overwrites the previous one, or is copied to a distinct location, such as a dated folder. The advantage of the latter approach is that you can roll-back to previous backups if newer ones are corrupted or no longer contain recently deleted files.

Full backups are sometimes preferred because of their cleanliness and simplicity. They are a precise clone of a data set which - unlike other backup methods - don’t rely on specialist software to function. You can take away a full backup, open it on another computer, and access the data instantly.

Restoring data from a full backup is relatively simple. Because it’s an exact copy, you can easily navigate to desired files and folders, restoring individual files on demand. Or, you can simply copy and paste the entire data set back to its original location. Of course, data added or changed since the last full backup will be lost after a restore.

The problem with full backups is they take time. 60GB of data could take an hour or longer to backup, and the process could disturb business operations or force you to schedule backups out-of-hours. Such pressures might prevent you from doing them as frequently as you should, exposing you to the risk of significant data loss.

Incremental backups

Incremental backups begin with a full backup. Once that’s done, only amended or newly created data is subsequently backed up. The result: a full backup plus a collection of new data sets containing new/amended data.

The process requires specialist software, which identifies changed or added data and manages the backup and restore processes automatically. Some incremental backups keep archives of previous versions of amended/changed files so that data can be rolled back to a previous state. The backed up data may not be accessible manually, or accessed on another computer without installing the backup software.

The advantage of incremental backups is they are usually much quicker than full backups. The exact duration of an incremental backup depends on the amount of data added or changed, but typically they will take a fraction of the time needed to conduct a full backup. Some modern backup programmes are also more capable of conducting on-the-fly backups, which means you might not need to halt data activity during backups.

Because incremental backups (and subsequent restores) are dependent on specialist software to function, it’s vital to select reliable and robust backup software. The process of working out what data to restore is a complex task, which in rare cases could fail. Modern software is becoming much better at the task, but it’s important to test programmes thoroughly and have confidence in their reliability.

Note: If you’re unsure about which backup approach to adopt, remember: you can use both independently. To illustrate, you could manage a weekly cycle of full backups on one external hard drive, and implement an incremental backup process on another. If the latter process fails you could revert to the previous full backup… The most important thing is to consider the strengths and weaknesses of each method. If the weaknesses of either approach amount to an unacceptable level of risk, consider how a combination of different approaches could be used. (That said, it’s also easy to overcomplicate things - your backup strategy needs to be robust, but also simple enough that you’ll keep it up.)

Backup Mediums

Tape. Suited for large-scale, ’set-and-forget’ automated backups. Tapes are relatively inexpensive but tape drives are costly and relatively slow. Tapes are durable and compact, so they can be easily moved, stored and archived on or off site.

CD/DVD. Cheap, compact and portable. No moving parts, so when in good condition they are relatively stable. Disk capacities are a fraction of modern hard-drives so full backups may require numerous disks, and the process must be managed manually or with backup software. Because disks may need to be changed they are not ideal for unattended/automated backups.

External hard drive. Modern USB and Firewire drives are fast, reliable, and offer storage in excess of internal drives, making full and incremental backups easy to manage. Their portability means they can be easily transported off-site, but like any hard drive they are sensitive to shock, heat and moisture. Some external hard drives support “RAID” configurations for simultaneous mirror backups over multiple hard drives. Some drives can also be networked. All hard drives contain sensitive moving parts, so external drives are as (if not more) susceptible to failure as internal ones.

Server. Servers can be used as a backup location for data stored on individual computers, or they could be employed as a ’shared drive’ upon which users store their data in the first place, which is then backed up onto another backup medium. They are expensive, but often support advanced features such as “RAID” which allows automatic ‘mirroring’ of data onto another server hard drive, stored on or offsite.

Online ‘cloud’ storage. Online storage can be used as a networked hard drive; meaning you can drag or drop files like an internal/external drive, or use your own automatic backup software. Or, they are designed as proprietary backup services in themselves; usually incremental backups which can be done automatically whenever you are online, at a pre-determined time, or whenever your computer is idle. Capacity is usually modest (at a fraction of a typical internal hard drive) but most services offer increased capacities at a cost. A fast, reliable internet connection is crucial if regularly transferring large amounts of data. As with off line incremental backup services, the reliability of the backup process is paramount to the ability to successfully restore. On top of that, online services must be chosen carefully, considering factors such as: guaranteed ‘uptime’ of service, safety and security of where your data is stored, how quickly restores can be accessed and downloaded.

Piecing together a strategy

Rule number one: the more backups in the chain, the lower the chances of losing everything… Rule number two: using different backup approaches and backup mediums helps minimise the weaknesses of one single approach.

Paranoia can be a good third rule when it comes to backups. And it’s easy to become paranoid when you contemplate losing everything. But worrying too much could lead to obsessive backup processes that are overly costly and time-consuming. It’s therefore important to find a balanced backup strategy that reflects the value of your data (or put another way: the cost of losing your data), but one which is also pragmatic about what you can afford to do, and what you can (realistically) be bothered to do.

When defining a backup strategy remember the following… If your data is invaluable, be bothered… The more backups the better, within reason… If you’re not 100% positive on one approach, consider others ways to crack the egg… All backup mediums - regardless of cost - are prone to failure, so consider the strengths and weaknesses of each and use a combination to reduce overall risk… Always have an off-site backup in case of fire or theft… And finally: settle on a backup strategy that protects your data as far as you reasonably can, but one which is realistic enough for you to keep up.

More info

Check out online data backup and storage services in our list of cloud applications and services.