I was asked during a disaster recovery session if the cloud could be used as a destination for backups. I guess this question could be simply answered as yes. I’m sure you already know that storage in the cloud is being used for data backup and on some scale for disaster recovery. While I can say yes to the question about “Can backups be sent to the cloud?”, the bigger question for me is “Is it ready for large scale backups to the cloud that can be used for disaster recovery?”. I’d also ask whether backing up data to the cloud is the right thing to do giving the requirements. You really have to look at the big picture here right. Disaster recovery has many layers and tiers with many solutions to fit those layers or tiers based on many things like Recovery Point Objective and Recovery Time Object. It may not have to be complicated for small environments but I’m talking about dealing with terabytes of data here, not the free storage space you can get from Dropbox or Box.net that you use to backup your family photos.
I have been seeing a growing number of companies hopping on the bandwagon to support or solution backups to the cloud. Not that this is a bad thing or area of focus if used in the right way but I don’t see it as the light bulb turning on over my head “ah ha” moment. Just looking at the cloud as a place to copy data so that when that time of crisis happens there is a highly available copy which can be recovered is in my opinion not looking at the big picture. Here is how I think the cloud should be looked at before you start sending backup data to it.
First you need to identify your storage and I’ll keep it simple here. Label your storage into a couple different areas and analyze how it’s used day to day in your environment.
- File Shares – typical smb share that user access and store files
- Content Management – A system that stores and manages files like MS Sharepoint
- Application/Database – Varies applications that use a database backend to store data or information
These are just three areas in which data is stored and can consists of many servers in your environment. I can’t tell you how many redundant systems I see in a single environment that do the same thing but are separated for some reason but that’s a different topic. Once we, at a high level, label these different types of data we can see that some of these systems could actually live in the cloud. One point I’m trying to make here is that maybe you should be looking at moving the application into the cloud instead of just its backups. If the servers that host all of this data is in the already in the cloud you remove issues like how to get the backup data of those servers quickly and securely into the cloud. And the more important issue of how do you pull that data back into your network quickly and reliably in order to restore services is removed. Just think about it, in order to restore the data you’ll also have to have infrastructure to do it with.I think it would be a bigger nightmare using the cloud to backup data in old traditional backup to disk method. A better way is to just move your data and services into the cloud where possible. In a lot of cases you could improve the service or access to data, add redundancy, and recover faster. So if your planning on using the cloud for the purpose of disaster recovery or to store backup data it maybe wise to get more than just a cloud archive. And use data protection and disaster avoidance as part of your justification for moving applications into the cloud.