The three major data backup methods are: full backup, incremental backup and differential backup. Through simple analysis, I hope to improve everyone’s understanding of storage backup.

Full backup of data backup method

The backup system does not check whether the files have been changed since the last backup; it just reads and writes each file mechanically, regardless of whether the file has been modified. Back up all selected files and folders, do not rely on the file’s save attribute to determine which files to back up.

This is why we don’t just take a full backup — every file will be written to the backup device. This means that even if all files are unchanged, they will still occupy a lot of storage space. If the files that change every day are only 10 MB, and you need to spend 100 GB of storage space every night for backup, this is definitely not a good method; this is the main reason for the introduction of “incremental backups”.

Incremental backup of data backup method

Unlike full backups, incremental backups will first determine whether the last modification time of the file is later than the time of the last backup before data backup. If it is not, it means that the file has not been changed since the last backup, so no backup is required this time. In other words, if the modification date “does” come later than the date of the last change, the file is changed and needs to be backed up.

Incremental backups are often combined with full backups (such as full backups every week and incremental backups every day). Differential backups are for full backups: back up all files that have changed since the last full backup.

The biggest advantage of using incremental backup is the backup speed: it is much faster than the full backup. At the same time, because the incremental backup will automatically determine the backup time and whether the file has been changed before making the backup, it is Saving storage space is also beneficial. The disadvantage of incremental backup is that it takes a long time to restore the data and the efficiency is relatively low. For example, if you want to restore a backup file, you must search all the disks of the incremental backup until you find it. If you want to To restore the entire file system, you must first restore the most recent full backup, and then restore one incremental backup after another.

For example:
If the system fails on Thursday morning and loses a large amount of data, then it is now necessary to restore the system to the state of Wednesday night. At this time, the administrator needs to first find the Monday full backup tape for system recovery, then find the Tuesday tape to restore the data on Tuesday, and then find the Wednesday tape to restore the data on Wednesday. Obviously this is much more troublesome than the first strategy. In addition, this kind of backup reliability is also poor. Under this kind of backup, the relationship between the tapes is like a chain, one by one, and any of the tapes will cause the entire chain to be disconnected.
The most obvious advantage of this backup method is that there is no duplicate backup data, so the amount of backup data is not large, and the time required for backup is very short. But the data recovery of incremental backup is more troublesome. You must have the last full backup and all incremental backup tapes (once one of the tapes is lost or damaged, it will cause recovery failure), and they must be reversed one by one along the chronological order from full backup to successive incremental backup. Push recovery, so this greatly extends the recovery time.

To avoid restoring incremental data one after another, to improve the efficiency of data recovery, and change the method a little, it becomes a “differential backup (differential backup)”.

Differential backup of data backup methods

Like incremental backups, differential backups only back up changed data. But the backup of the former is “cumulative”-as long as a file has been updated since the last full backup, then every time you make a differential backup, this file will be backed up (of course, until the next time Full backup).

This means that the files in the differential backup are all files that have been changed since the last full backup. If you want to restore the entire system, you only need to restore the full backup first, and then restore the last differential backup. Incremental backup is for the last backup (no matter what kind of backup): after the last backup, all the files that have changed.

Just like the strategy used for incremental backups, you usually only need to make a full backup regularly, and then do differential backups regularly.

Therefore, the size of the differential backup will continue to increase over time (assuming that the files that are modified every day are different between full backups). In terms of backup space and speed, differential backup is between incremental backup and full backup; but whether it is to restore a file or the entire system, the speed is usually faster than full backup and incremental backup (because the number of disks to search / restore a bit less).

For example , on Monday, the network administrator routinely performs a complete system backup; on Tuesday, it is assumed that there is only one more asset list in the system, so the administrator only needs to back up this asset list together; on Wednesday There is another product catalog in the system, so the administrator not only needs to back up this catalog, but also the asset list on Tuesday. If there is an additional salary table in the system on Thursday, the content to be backed up on Thursday is: salary table + product catalog + asset list.

Thanks for reading this.