Part of your data disaster recovery plan should include the types of backup you intend to apply to your data. Note if you depend on only one type, backing uo to one location you may well regret taking the easy solution.
You might be asking “What are the different types of data backup solutions” – The following is little more than a glossary, but it should help to understand the purpose of different backup types and which of those are best suited to your business model.
You should be able to use this knowledge to help choose the most appropriate, hardware, software and offsite services that will offer you the best data security that matches your budget.
Hardware, Software, and Online Solutions
For information and recommendations on the hardware and software to use as part of your backup strategy check out the other pages in our Onsite section. For online or Cloud Backup check our offline section.
Backup Schedules / Timing
This method is usually used to save all the user data from a specific disk (usually not the OS). This type of backup is likely to be performed on a specified schedule, maybe every day or every week.
This method backs up only the data that has been modified since the last full backup. So if you do a Full Backup every Friday and a differential backup every day, if your data is lost on a Thursday you first restore the full backup then Wednesdays data. You will still have lost Thursdays data. Differential file sizes increase in size each day.
This method backs up only the data that has been modified since any a previous backup (Full, Differential or Incremental). Incremental backing up of files is often performed every day or more frequently. If you only perform Full and Incremental backups then if your data fails on a Thursday you will need to restore the last Full Backup then every incremental backup in order since the last full backup. Incremental backups are the smallest in size, thus use the least storage space and fastest backup – but can take a longer to restore (with more opportunity for error or failure) if you do not also use a Differential backup.
Some organisations would do a full backup every week a Differential backup every day and an incremental backup every hour. This offers the best compromise between effort and resources required for backups and offers the best and most efficient and successful restoration (assuming some real time mirror backup is not in place)
Real Time Data Mirroring
This strategy copies all changes instantly, as they happen. In fact it is usually not a copy process, when data is written or updated the file system simultaneously updates to both devices (usually hard disks) Because data is duplicated in real time, the data in both locations / devices is always an exact copy. Data mirroring offers nearly instant recovery if one device fails or one location is hit by some major physical or data disaster. Real Time mirroring, particularly if each data store is in a different location can require more expensive infrastructure and data processing and transmission costs.
Hot backup / Dynamic backup / Active backup
When using this system data is backed up when it is currently accessible to users. Hot backups are a very common in multi-user systems as there is no downtime that ‘cold backup’ methods require. Hot backups involve risks. If the data is being altered by a user during the process the final version may not reflect those changes. To reduce this risk a redo log will be created at the during the backup. If restoration is required the log will identify files that may require further verification before the restoring.
Automatic backup system
This system is triggered by some event—for example, a schedule point or a threshold reached—rather than by human action.
What data is backed up – in what format
Bare-metal backup and Restore
A bare-metal restore (BMR) is a method where the full system, OS, programs and user data can be restored (to a new empty device if necessary) without the need for the original OS or program disks. This system is different from a full backup which normally only backs up user data – not the OS.
With file backup any file that has had any change since the last backup will be backed up in full – even it that is a very large text document and only one word has been changed. This method tends to be slower than block backup however it offers more flexibility and speed if only a few individual files need restored.
Block-level ignores the actual files stored on the system, but rather analyses blocks of data and backs up those blocks that have been changed since the last backup. The size of the blocks can be adjusted within the software to be large or small blocks. The smaller the block the slower data is backed up.
These blocks are just blocks of data with no direct relationship to actual files – a block could be a part of one file or could contain parts form several files. Block is one of the fastest systems for backing up and for restoration but is less flexible when individual files need restored.
For some more technical information see this page on the University of Maryland website.