Image by craig1black via Flickr(I have borrowed heavily from Wikipedia as a reference for the terms at use today. In large part , I believe that much of this article has touched on “common knowledge” (ie: there wasn’t much new material here from what we discussed as a class). And while I have used formatting similar to the current Wikipedia article, I have tried to use my own language in describing otherwise common concepts.)

Backup is a secondary copy of data (whereas ‘archives’ are the primary copy). Backup can be employed in primary two – though necessarily exclusive – ways. One way is to keep a complete, full system backup (everything on your computer – even down to your operating system, etc). Restoring such a backup would give you a full working system identical to the one the way it was running when the backup was made, with all accompanying data.

A second route is to backup data files only. In this way, the files can be restored in relatively small increments. This is a particularly useful approach when installation of an operating system is not particularly unique or difficult. It also typically requires less in resources and employs a number of techniques as listed below to make the backups quickly and efficiently.

Data repository models

• Unstructured: Keeping your backup files on cd/dvds in a pile under your desk – with little regard to what is included or when – is a haphazard approach to backup that never seems to satisfy the need. Believe me, as a recovering unstructured backup guy, there is little comfort in this approach.
• Full + Incremental: To clear the air, the definition of a ‘full’ backup is largely user defined (‘full’ then meaning 100% of the files that the user wants backed up – whether it is a full system backup or a set of data files.) Thus, after a full backup is taken, this method then backs up all files new files and files that have changed since the last full or incremental backup. One advantage to incremental backup is that it is usually faster to complete the backup cycle. A downside to this approach is that one needs the entire chain of incremental backups to make the full backup in the case of a full restore.
• Full + Differential: Differential is similar to incremental in that it targets the new and changed files since the full backup. It differs, however, by creating a backup that contains all new or changed files since the full backup. So there would be two backups (the full and the differential), instead of many (full and the many incremental).
• Continuous Data Protection: These backups are done the instant that a change is found in a file. This is typically done by saving the differences in bytes rather than at the file-level (as the file may still be open, etc).

There are many types of media that are employed in backup. These include:

• Magnetic Tape: Primarily a corporate medium, with a notable absence in the consumer market. Tapes are relatively inexpensive and have proven t be a dependable medium. Like the music cassettes of yesteryear, the access time to information on the tape will be slow (sequentially recorded). However, remember “high dub”? – tapes can record at speeds near hard drives.
• Hard Disk: High capacity and low access time. Hard disks make it easy to put large amounts of data and/or system images into storage at relatively high write speeds. Local connections like SCSI, Firewire, or USB provide standard connection interfaces. Networked connections through ethernet, fiber, etc provide more opportunities for use. Hard disks have demonstrated sustained growth in capacity and declining pricing on the technology over the course of the past two decades.
• Optical Disc: I remember a time when everyone who was anyone was out buying a cd-writer to, among other uses, backup their data. (And how could we ever fill 700mb! Whoa!) DVDs followed with a 4.7gb capacity – and subsequently released a double layer DVD that held 9.4gb. Using optical media for backup worked very well for people for a number of years because of the relatively high capacity of the media, the relatively low cost of the media, and the portable nature of the media. The media could be taken off-site for added security against theft or natural disaster/fire, etc. There has been a lag in optical media over the past few years as capacities have failed to rise in the face of low cost flash drives and hard disks that offered greater capacities and read/write ease. Only in the past few months has Blu-ray discs – with a capacity of around 50gb – begun to take a foothold in the industry.
• Floppy Disk: I recall sitting in my junior high class (and even high school) and being required to bring a floppy disk to class to save homework on. The capacity of those disks was somewhere around 1.44mb. Yeah… The first flash drives that came to market completely blew these disks to oblivion. Curiously, my ward on campus uses these same floppy disks to back up financial data.
• Solid State: This encompasses a wide range of implementations including flash drives (or thumb drives, etc) that have been referenced previously. These tend to be the portable tool of choice for student and professional alike in current day. Unfortunately, reliance on these devices alone for backup tend to be the fall of many. Either through loss of the devices or corruption of files, many a student has lost a semester’s work (usually during finals week). Memory cards such as Secure Digital, Compact Flash, etc have also enjoyed widespread adoption in the digital camera market (and throughout consumer electronics in general). Solid State hard drives have also begun emerging as an as of yet pricey alternative to hard disk for laptops.
• Remote Backup Service: Many companies have begun offering backup services in their data centers (usually hard disk) over the internet. Advantages are that the off-site storage protects against theft, fire, and other natural disasters. Disadvantages include relatively slower speed and putting your data in the hands of a third party.