Data loss is a known peril. When dedicated server and VPN customers lose their website data, they stand to lose their livelihood or precious memories. Customer support teams deal with major or minor issues day in and day out but no amount of hard work wins them more gratitude than a flawless crash recovery. It may feel like a miracle for new customers but due the way data backup is handled in the industry, it’s a simple 2-3 step process. In some cases, like automated cancellation backups are retained for 45-90 days because companies expect customers to come back!
Businesses need reliable backups to ensue business continuity. They can bounce back from an unscheduled downtime but there’s no recovering from data loss. Today, customers have dozens of alternatives, even the most loyal ones move on as soon as you’ve lost their data. Therefore, backup is one of the most critical aspects of keeping an online business afloat.
What’s a backup?
Backup in the web hosting industry refers to the process of making scheduled copies of a website or server’s data. It may be a mirror image or a backup of mission critical data like email and database. Since operating system and other software can be reinstalled and customized settings can be saved in a file, some admins choose to remove these files from backups to reduce the size.
Types of backup
Historically, there have been three types of backup, and they are still in use depending on requirements of a business:
Full backup: There are no checks, in a full backup all files are copied from source and pasted to destination. Every single file in an account or server is backed up. It sounds perfect but the disadvantage is using up a lot of disk space over a period of time. Reliable disk space is expensive. Since backups are made hourly or every day, it is easy to run out of 100 GB space.
Incremental data backup: Incremental backups start with a full backup but every subsequent backup checks for changes using a reference point. Files that have not been modified since the last backup are skipped. If a file has been modified since the reference point, it is added to the current backup. Some backup systems can go through all the incremental backups to locate the changes and generate up to date restoration data.
Differential data backups: Differential backup is similar to incremental backup and saves the data that has changed since last backup. Its advantage is that it needs only two data sets for a full restore. It takes the most recent full backup and applies a differential backup to it. However, the downside is that the time to restore increases depending on the number of changes between the two data sets.
Mirror Image: A mirror image backup stores a recent “mirror” of the source data at different intervals. It starts with a full data backup and maintains a live copy by using hard links or binary diffs. It is particularly effective for large backups that don’t change frequently. The Mac OS time machine is an example.
Continuous data protection: The continuous data protection process involves logging all the changes to the source in real time. It saves byte level differences unlike traditional backups that are based on file level differences. It is also different from mirroring in the way that it can analyze the logs and restore an image.