User login

Disaster Recovery

While it is an extreme case about the importance of a sound data recovery strategy, in January of 2011, a cancer researcher at Oklahoma University had a laptop stolen by a thief smashing a car window. The white MacBook had five years of work on it. The reasearcher believed she was on the verge of a cure for prostate cancer; the information contained in the laptop was invaluable in her research. But lacking a backup of any of the data on the laptop, five years worth of research disappeared in a flash. This cancer researcher's case mimics the case of hundreds of thousands of disaster recover scenarios that occur each year worldwide. It is just the scale of the tragedy that makes it an important example. While you may not have a cure for cancer, if your laptop was stolen today, or your desktop system hard disk crashed, or there was a fire in your server room, and none of these assets were properly backed up, then you would likely come close to feeling what that researcher felt. For others, the equivalent may be the loss of your accounting database or the loss of all of your personal photos or the loss of every email you've ever received.

It is very easy to become complacent. On the laptop this article is being written, it has been in service for 2-1/2 years without a single event that would cause a scare. No misplacement, no hard drive crashes, no other issues which have hindered its service for more than a few minutes at a time. It is luck, pure and simple. It gets easy to forget that something could happen to it at any moment. But it gets backed up every day. It scares the hell out of us not to back it up and every other computer or server, or cell phone, or tablet in the enviroment. Its demise, whether lost, stolen, or crashed, becomes a mere annoyance rather than a potential nightmare. Between 256-bit hard drive encryption and having two complete, spatially-separated backups -- in the event of its loss or theft, the peace of mind that some hacker is not stealing our stuff and we can recover very quickly is very satisfying.

But if you have a server with just 250GB those systems become impractical to back up to the cloud. Considering a simple wideband connection (DSL or T1) you may be able to upload as much as 2GB of data per day, so getting a backup of 250GB would take 125 days to accomplish. Even 30GB in the example above might take 15 days to get the initial image. But let's assume you have ten times the download bandwidth than upload bandwidth and have to move 250MB back down into a new or repaired server. It still might take 12 days of download time to get your server back to 100%. An online backup strategy for a server with 250GB would also cost in excess of $100/month in storage fees from a service like Mozy. Removable drives, or even cheap USB backup drives are fairly cheap solutions (often less than $70 per drive). With hot swappable hard drives, moving a copy offsite takes just a few minutes and a walk to the car. Encyption ensures your data is safe if the drive is stolen before it arrives at its offsite location.

Carrying a copy of a desktop backup offsite and keeping it separate from the computer exponentially increases the probabitly of recovery in a disaster. Mapping multiple computers and backing up to a simple drive makes a small, server-less office a single drive solution of two drives in constant rotation. The rule of thumb is to have three copies at the very least. The computer being the first, a local backup, and an offsite backup. An investment of a couple of hundred dollars can save tens of thousands or more in lost data.