Archival Storage: Routines, Software & Hardware

Why we bother

Computers crash. Hard drives die. Laptops are stolen. With a proper backup routine, these inevitabilities will make for a bad day. Without a proper backup routine they can ruin years of work.

Many photographers feel that digital photography lacks the ability to archive images as well as film. This is just not true; with proper discipline, a digital image can be better protected against aging and catastrophic loss than any piece of film. A digital image can be perfectly copied, again and again, without loss. However, if backups are not done properly an image, or even years of images can be lost forever in seconds.

Remember: There are only two kinds of data: data that is backed up, and data that hasn’t been lost yet.

Jump to

• Backup Routines and Methods
• Hardware Solutions
• Software Solutions
• RAID: Good or Evil

_

_

Archival Storage Routines and Methods

Primary+On-Site-Backup(s)

This routine calls for two or more copies of all of your data: the primary copy and one or more backups which remain in the same place as the primary copy. This will protect against most types of loss, but will fail if there is catastrophic loss (e.g. fire/flood/vandalism).

Primary+OnsiteBackup+OffsiteBackup

This routine calls for keeping three copies of all of your data: the primary copy, an on-site backup, and an off-site backup. Every day the primary is copied to the on-site backup. Every month the on-site backup is swapped with the off-site backup. The three copies will never be in one location at the same time, so in the case of catastrophic loss (e.g. fire/flood/vandalism) at the main location there is a safe copy of the data.

_

Hardware Solutions

It appears that Apple may be discontinuing the XserveRAID in favor of promoting Promise brand RAID enclosures. More information will be posted when it becomes available.

Very High-End Solution: Custom RAID enclosure

This is non brand specific. Each situation requires slightly different capabilities, and at this level (very high-end) we’ll match the specific requirements to the best brand.

• Call for price estimate including installation and setup
• best for 2-10  terabytes of ultra-secure triple-redundancy backup when used with the primary+on-site+off-site method
• Redundant power supplies w/ Cache battery backup
• Redundant RAID controllers
• Redundant Cooling System
• Warranty Service provided at Apple Stores
• Raid 5 with hot-spare compatible
• Fibre connection to G5 or MacPro
• Much faster than an internal drive and light-years faster than a firewire drive

High-End Solution: Two Drobos & twelve 750gig drives

Four Primary drives, four On-Site-Backup drives, four Off-Site-Backup drives

• Retail price: $499 for each Drobo not including drives
• 2.1 Terabytes (2,160 GB) of usable double redundant storage
• Drobo automatically combines up to four hard drives into one pool of protected storage
• Extremely easy to use:
  • Green lights mean your data is safe even if a hard drive fails.
  • Yellow means your data is safe, but you should add/replace a hard drive soon.
  • Red means your data is not safe and you should add a drive immediately
• Drobo can mix and match hard drives of different sizes

Suggested configuration: one Drobo on-site, one Drobo off-site; three sets of drives (primary, on-site backup, off-site backup) four drives each. Each set will have 1.8 terabytes (1,800GB)of usable space.

Entry-Level Solution: One Drobo & three 500gig drives

Three primary drives, three On-Site-Backup drives, three Off-Site-Backup drives

• Retail price: $499 for the Drobo not including drives
• 1.3 Terabytes (1,260GB) of usable redundant storage
• Easily expand capacity in the future by replacing a 500 gig hard drive with a 1TB hard drive.

_

_

Software Solutions

Feature	Time Machine	SuperDuper!
Price	Free with OSX 10.5 Leopard	$27.95
Automation	Every Hour	Every Day
Incremental	Yes	No
Delta	Yes	Yes
Bootable	No	Yes
Notes	Cmd Click on icon for manual backup	www.shirt-pocket.com/SuperDuper/

Automation

The amount of work it takes for the user to create and keep a backup is critical. It’s less likely that a backup will be kept up-to-date if the software doesn’t offer some degree of automation.

Incremental Backup

Incremental backups keep multiple versions of any given file. If your assistant accidentally erases or overwrites a file, an incremental backup will contain the file before and after deletion/overwriting, allowing you to correct the error. Most incremental backup programs compare the existing data to the previous backup and ignores any data that hasn’t changed since the last backup, saving time and space.

Delta Backup

A delta backup, also called “Smart backup” and “dynamic backup” by different companies, saves time during a backup by only copying those files which have changed since the last backup.

Bootable

A bootable backup is just that: bootable. If you keep a bootable backup of your laptop hard drive and the hard drive in your laptop fails, then you can physically swap the dead drive with the backup and keep going without any need to reinstall the operating system or programs. Better yet, plug in the bootable backup via firewire, and power on the computer while holding option/alt; you’ll be able to select to boot from the external hard drive instead of the failed laptop hard drive.

For digital techs, a firewire-powered external bootable backup with OSX, Capture One, and Photoshop is a great tool to have. If there are software problems or hard drive problems, boot from your external backup and keep shooting.

_

RAID: Good or Evil?

A RAID takes two or more physical hard drives and combines them into one virtual hard drive. There are many types of RAID; here are the most relevant to photography:

RAID0 (a.k.a. “stripping”)

RAID0 splits every file up into parts; if there are two drives in the array then each drive will get half of every file. The benefit is speed since each drive has less to record. If any of the drives in the RAID fail all of the data is lost. This is NOT a good idea in photography since it increases the chance of data loss.

RAID1 (a.k.a. “mirroring”)

RAID1 stores every file on every drive in the array. At every moment, each drive in the array is a perfect duplicate. The benefit is that if one drive fails, there is no loss of data.

RAID5 (a.k.a. “parity”)

RAID5 is harder to understand than RAID0 or RAID1. Here is an analogy: imagine you assign four people to remember three numbers, say “7, 3, 5″. In RAID5, the first person would remember “7″, the second, “3″, the third, “5″, and the fourth person, who we’ll call Mr. Parity would remember the sum of the numbers was “15″. Now if the first person forgets their number then Mr. Parity and the two who remember their numbers can figure out the missing number (3 + 5 + ? = 15).

RAID5 is great because it offers protection against drive loss, and because it lets you combine several hard drives into one big hard drive. Instead of having to divide your photo library and keep track of what is where, your entire library can fit on one array.

Why RAID isn’t a backup by itself

• Data that is deleted/overwritten/corrupted on one drive is instantly deleted/overwritten/corrupted on all the others
• The RAID controller itself (hardware or software) can malfunction and corrupt the data on all the drives. This is rare, but it does happen.
• All the drives have to be physically located together; in the case of fire/flood/vandalism, all the data is lost.

Confused Yet? Get a Drobo.

If you’ve read this section on RAID and you are still confused, you are not alone. RAID can be difficult to understand, and tedious to setup and maintain. That’s why we suggest Drobo, the storage robot. It has most of the advantageous of RAID, but without the limitations and complexity!