Cloud-based disaster recovery has a lot going for it, including cost reduction for hardware, software, maintenance, and staffing. Not only can organisations avoid the large initial investments associated with traditional DR and the need to buy their own servers, data centre space, etc., they can also restrict much of their ongoing costs to what they use. As IT disasters hopefully will be few and far between, use of cloud-based DR should be modest and costs low. However, this utopic vision may be marred by several factors that can cripple cloud DR performance, destroy end-user confidence, and, yes, even make cloud DR more expensive than its traditional counterpart.
- The first challenge comes from accessing business applications that have been moved or replicated off premises and up in the cloud. If applications have not been designed to run for remote instead of only local users, network bandwidth can easily be used up. Possible solutions include the use of virtual desktop software to limit the amount of data zinging to and fro between user devices and the application.
- Security is the next item to be addressed, both to allay fears of some end-users that their data may not be safe “up there in the cloud”, and to remedy carelessness of others whose security precautions are inadequate. Encryption of data at rest, in transit, and in processing can play an important role in dealing with both these security aspects.
- Although cloud usage is “pay as you go” with little or no initial expense, excessive use of cloud resources can leave you with a big bill. Making sure disaster recovery facilities can be quickly spun up as virtual machines in the cloud, rather than leaving them on permanently, sounds like common sense, but what about the cost of transferring data for your DR cloud facility to work with? If you work with terabytes of data and on-premises servers, the price of sending those terabytes over the network and up into the cloud can be astronomical – not to mention the time it takes. Plan, and make incremental data backups where possible, rather than trying to transfer all your corporate data in one go in an emergency.