Business Data Backup Methodologies, Solutions

Executive Overview

Computer system backups are a critical part of the compute environment. Data loss can cripple a business. Reliable, tested and proven backups are a necessity that may not have apparent short-term value but will ultimately prove invaluable.

The measured cost of backups, both in terms of capital equipment outlay and ongoing consumables, must be weighed against the cost of a potential catastrophic data failure. It is difficult for us to assess the cost of restoring data that cannot be recovered from backup. More than one business has been wiped out due to catastrophic data loss and irretrievable backups.

A backup plan consists of four parts:

Decide what to backup
Decide how to backup
Schedule and test the backup
Schedule and test recovery procedures

Arguably, step 1 is the most important. There are considerations beyond the “back up everything” mind set. Different storage options have different capacities. Backup time is critical—if backups take too long, they become a detrimental inconvenience and potentially run in to “eat your own tail” situations.

We will deal with two “how to” scenarios in this document:

Local backup to different media
1. Disk-to-disk
2. Disk-to-tape
3. Disk-to-disk-to-tape
Remote, offsite backup

There’s a simple maxim that describes backup: “Nobody cares if you can back up. They only care if you can restore.”

Backup Scenarios

Type	Relative Cost	Relative Complexity	Pros	Cons
Disk-to-Disk	Low	Simple	Backup is fast Limited only by disk size Recovery can be fast	Subject to catastrophic failure. Disaster recovery best practices requires multiple disks, at least one stored offsite, which requires management Generational backups require planning.
Disk-to-Tape	High	Complex	Backup can be fast, depending on media Limitless capacity—simply buy more tapes Recovery can be fast, depending on media	Consumable costs can be high. Tape library requires management. Distaster recovery best practices require tapes to be stored offsite, which could delay recovery. Generational backups require planning.
Disk-to-Disk-to-Tape	High	Complex	Fastest solution Limitless capacity—simply buy more tape. Disks acts as a “cache” before data goes to tape, boosting performance.	Consumable costs can be high. Tape library requires management. Distaster recovery best practices require tapes to be stored offsite, which could delay recovery. Generational backups require planning.
Remote Backup	Middle	Simple	Backup can be constant and ongoing—files get backed up as they change. Very little management beyond initial setup. Data is kept off site and can be used to rebuild a system in the event of a catastrophic failure. Generational backups are automatic.	Costs can be high depending on amount of data to protect Performance is dependent on bandwidth. Initial setup takes time.

Disk-to-Disk

To clarify: disk-to-disk backup refers to copying data to an external or removable disk. Disk-to-disk backup to another internal, non-removable drive is not a viable option.

Disk-to-disk refers to using externally mounted or internal bay removable drives to hold the backup. Multiple disks are required to ensure the backup integrity. At set intervals, backup disks are rotated out.

The capacity of the external disk is determined by the size of the backup set and needs to be calculated to include growth. In addition, factors such as “generational” data must be considered.

The number of external drives is determined by the frequency at which the disks get rotated out. The longer a disk is in use, the greater the potential data loss in the event of a datacentre-level catastrophe. If you backup a month’s worth of data and the data centre burns down on day 29, you’ve lost an entire month worth of data since you’d only be able to restore from the previous months disk. In addition, if you have too small a number of drives, your data retention becomes limited. Two drives rotated monthly, for example, mean that you only have 60 days of data retained.

Disk-to-Tape

Data retention and data volatility issues can be overcome by backing up to tape. Tapes can be rotated as frequently as required. Tape capacity is high depending on the media type. Tape changers can automatically rotate tapes, essentially multiplying the capacity and making rotation management much easier. Transporting tapes offsite for safety is very easy.

There are two tradeoffs here: speed and cost. High capacity drives and tapes can be expensive. Tape changers can be expensive. Data retention policies can require a large number of tapes, which increases the consumable cost.

This method resolves the most pressing issues with Disk-to-Tape and combines Disk-to-Disk and Disk-to-Tape to get the best of both. Backup performance is high because data is streaming to disk. Data retention and volatility issues are met once the cached backup is spooled to tape. Tapes can easily be transported off-site for safe keeping.

The three preceding methods require software to manage the backup sets and things like data retention policies. On the server side, this software can be quite pricey.

Remote Backup

In the day of high bandwidth Internet connectivity, Remote Backup has become a viable alternative to more traditional methodologies. Data is streamed from a server to a data storage “cloud”. Services like Mozy encrypt the data before it is sent to the cloud, so data safety is assured. Backups are constantly available so recovery is simpler than any other solution. Costs are relatively low—protecting 100GB costs as little as $60/month, which is often less than the cost of consumables when running a tape drive.

Initial setup of remote storage can take considerable time depending on the size of the backup set. Once the initial image has been taken, though, many services offer ongoing real-time backup—any time a file is changed, it’s automatically pulled into the backup set in the cloud.

Different offsite vendors need to be evaluated to compare things like data retention policies, simplicity, cost, and performance. Most vendors offer similar feature sets, and most vendors provide the necessary server software at no cost.

Cost Comparisons

Type	Capital Outlay	Annual Costs
Disk-to-Disk	$300-$1200	$300-$12001
Disk-to-Tape	$18002	$24003
Disk-to-Disk-to-Tape	$2100	$2400
Remote Backup	<$100	$500-$?4

1 Depends on the number of disks used, which depends on the data retention and volatility policies.

2 Based on LTO 3 and enough tapes to do a simple 3 generation backup set

3 Consumable costs to cover quarterly retention and tape rotation.

4 Depends on size of backup set. Typical costs are $500/100GB.

- Adelaide Business Computer Services