It's a very difficult situation when I must tell a new client that they have to spend a minimum of $1,500 just to get to the point they were at hours ago. That's for a seven day turn around. Next day service starts at $12,000. Not new equipment, just a business' data recovery. Finally, that's assuming that the wizards at my favorite data recovery center can actually recover the data, which isn't guaranteed.
While the equipment that is used to access data is easily replaced, often the data itself is not. Furthermore, a business' reputation is often at stake. Needless to say, a critical part of your IT services is to plan for data loss. Let's look at the largest causes of data loss, along with the data backup philosophy that fits best with minimizing it:
Intentional and unintentional actions:
An end user accidentally deletes a file. Another forgets to save it in the first place. These are intentional and unintentional actions. Not malicious, but damaging all the same.
This is why applications such as Microsoft Office Suite, and now even Operating Systems, such as Apple's OS X Lion, offer automatic save technologies. Even with such technology, it is still easy to accidentally delete a file. Because of this, I am a fan of self-service local backups for end users. This way, they can recover from accidental deletions without needing to involve IT staff.
In the midwest, our primary concerns are fire, flood, and tornados. Other places might add hurricanes, earthquakes, and volcanoes.
An offsite backup is required for this type of data loss. Many small businesses rely on the "take a backup home" method to provide protection. However, they are still taking a risk that their home will be affected by the same disaster as their office. With the onset of cloud backups, it is possible to obtain inexpensive offsite backup that is thousands of miles away, allowing the greatest protection against disaster.
Theft, viruses, unauthorized intrusions all fit within this category. It's also the sad state of our world that we must acknowledge possible terrorism, and how to respond.
Like disasters, offsite backups are key, but another variable is also important. Offline backups. It's important that there is at least one copy of the data that the intruder cannot modify. This is especially important with cloud-based backup services, because they are essentially online all the time. To account for this, most services keep multiple copies of all files, and make it difficult to delete files and impossible to modify them.
This is the largest category with hardware failures, such as hard drive crashes, power failures and overloads, and data corruption. Cloud computing can have a business failure, which is where the service you are utilizing closes their doors, or changes their licensing in a way that is incompatible with your business.
Hardware failures often are dealt with by building in redundancy. Since hard drives are both a high point of failure, and inexpensive, IT staff will often place RAID (Redundant Array of Inexpensive Disks) arrays on servers, where downtime often means thousands of dollars of lost work. Virtualization, which I will cover in a later article, allows redundancy in actual computers, even further reducing the possibility of data loss, and even reducing the possibility of downtime.
Power failures and overloads are often accounted for with UPSs (uninterruptible power supplies), which perform two functions. First, they keep your equipment running during short power events. The other function they perform is that they notify the computer equipment of impending power failure so that the equipment can power down gracefully, reducing the possibility of data corruption.
Finally, the advent of cloud computing has amplified the ability for business failures to affect a company's data. Don't think of this as just a cloud business failing, but rather anything that could go wrong with an external service. Therefore, it is important to include the cloud service in your backup plans, or provide local redundancy if the cloud is the backup plan.
In a future article, we'll further discuss the strategies used, and how to balance them with the costs involved.
- Jon Thompson