Implementing a disaster recovery strategy is critical, but is often delayed for two reasons:
- It's complicated to evaluate business operations to find critical data that needs to be made available first after a disaster
- The perception that a disaster recovery solution that meets targets is too expensive
Both of these issues create friction that slows down adoption of disaster recovery strategies and technologies. As a result, IT stays focused on daily proverbial fires instead of actual fires, tornadoes, earthquakes, or floods.
Being able to recover quickly from a data disaster is more important than ever. The average yearly cost of downtime is $880,000 for mid-sized businesses, according to a report by the Aberdeen Group.
Whether your company has the budget for an online DR-as-a-Service solution, or wants to leverage existing tape backup hardware for DR – put a strategy in place before a storm or other emergency is on the local news.
The three main options that an SMB-sized company has when building a disaster recovery strategy are:
- Physically moving tapes or drives offsite.
- Replicating data between offices or to an offsite co-lo data center
- DR-as-a-Service from the cloud
This section will address the best situations for each solution, the disadvantages, and the always crucial – costs of each.
- Physically moving tapes or drives offsite.
Then, depending on the needs of your business, have another tape backup copy, for example each Friday of the month. They can stay onsite in a fireproof safe, for recovering an accidentally deleted folder, for instance. Most safes aren't melt proof, so after a certain amount of time the tapes will still be vulnerable.
Then, for archival purposes, make three monthly tapes, that can go to a bank safety deposit box, these are the last line of defense and should definitely be offsite.
The best situation for a tape-based DR strategy is when a company's RTO can be comfortably in the 2-5 day range. A retail business or a school that will be closed after a natural disaster, fire, or major theft is an example of an organization that can comfortably leverage offsite tape backups for DR.
The main disadvantage of tape-based DR is in day-to-day operations. The amount of effort it takes to replace a single accidentally deleted file or folder means that some user files just go unrecovered.
The cost of maintaining a tape DR plan is low when you consider the hardware has already been paid for, but it also creeps up over time. For example, it costs about $5,225 a year for weekly tape pick-up and drop-off according to a vendor survey using a sample environment with 2TB and 20 endpoints.
Using tape backups for disaster recovery can be pretty complex, but are a good way to leverage existing equipment to put a DR strategy in place.
- Replicating backup data between offices or to an offsite co-lo data center.
This is a popular choice for IT directors who have multiple offices with IT assets in each. There's a lot of flexibility for "roll-your-own" solutions with this approach, and it's possible to use existing hardware this way also. There are plenty of backup software options that will let you use old fileservers as a backup target and then replicate the backup server in the main office to one in a remote office.
Companies with data sizes above 5TB may also consider the "rep and break" tactic, by replicating data over the LAN between two arrays, then moving one of them to a remote location. This effectively "seeds" the remote copy.
Regardless of the method used, the key benefit is replication. In case of a disaster in the main office, you can VPN into the remote backup server and start to recover. This works just as well for production server crashes in the main office and helping users with their accidentally deleted folders.
The costs of this solution vary widely. If you happen to have multiple offices with existing IT assets then, just like above, it's possible to get disaster recovery ready with just a few extra licenses for the backup software. However, if you don't have a remote office, renting rack space in a local co-lo data center can get really expensive with the space, power, and management fees.
Alternatively, buying multiple backup appliances for different offices to replicate between can be a way to get strong data protection. Over a few TB, the price of multiple appliances starts to get pretty high, but some CFOs prefer a big up-front cost they can amortize, rather than a regular monthly bill.
- DR-as-a-Service from the cloud.
A cloud, or online, disaster recovery strategy has the most advantages for small IT teams with over 1TB of backup data that don't have time manage multiple data centers or tape rotation schedules. Double that for companies that have remote offices without their own IT assets.
To really use the cloud for DR the data has to be a replica of your live file system in its native format. Having a replica file system lets end users recover their own individual files over the web, and lets IT recover to dissimilar hardware, and have complete control over which files and databases are restored first.
One of the biggest advantages of cloud disaster recovery is the ability to instantly recover files, folders, and databases. However, the term "instantly" has two meanings in this context and it's important to be aware of each.
- Fast data transfer technology makes recovery feel instantaneous
Recovering an individual file from the cloud can seem to happen instantly, if the file is small and it's being recovered over an internet connection with 10MBps of bandwidth, for example. Recovering a whole server or entire office worth of data is a different story. Consumer-grade cloud backup services aren't designed for high-speed recovery, and will significantly extend recovery time.
A cloud disaster recovery service that uses the WebDAV protocol to transfer data over HTTP, just like the internet itself, will deliver significantly faster data transfer speeds. Using the WebDAV protocol allows communications between customer equipment and the data center to be done in a rapid multi-threaded mode, fully utilizing available bandwidth and shrinking backup and recovery time.
- Replicated file systems can be recovered instantly because data isn't kept in a proprietary backup format
Disaster recovery solutions that use a "snapshot and replication" strategy can be instantly recovered because data doesn't have to be unpacked, and IT can choose just what to recover, instead of having to download big chunks of data then pull the critical files out.
Nick Mueller is the corporate reporter at Zetta.net, a 3-in-1 online backup, disaster recovery, and archiving provider.