Q&A With Gartner's Don Feinberg on Database as a Service and Cloud DBs

Microsoft, IBM, Oracle and Sun are now fueling the growing fire around the database-as-a-service and cloud database markets, but what's the difference between these offerings and what's the appeal? Database guru Don Feinberg defines terms and raises important questions about reliability and security.
What are you more concerned about, having your data on a shared server (with virtual partitions) or having your data in the cloud on a shared server?

If it's or 1010data or Kognitio and they are serving you with multi-tenancy, you still know they are physically managing and securing that computer. In the cloud, you don't know where it is. If I buy two virtual machines, one might be in India and one might be in China. What happens if there's a civil uprising in China and the server that I'm running on happens to be there? Then I have a problem.

Today, I trust me first, I trust a hosted service second and I trust the cloud the least. I'd say it will be two to four years before the cloud gets to maturity. As they prove that they have the reliability and all the required capabilities for disaster recovery, then people will start to accept and trust the cloud.

What's the appeal until that time?

You're going to see it used for specific purposes. For a small company, it might be used as the infrastructure for everything. I don't foresee larger companies using it that way, but they will use it for development purposes and short-term projects. Instead of having to set up projects in a data center and having special equipment that may sit idle if nobody is developing, I can buy those services from the cloud, develop on whatever I want to use and then move it into my data center when it's ready for production use. I could also use it for short-term projects. I could set up a data mart for a one-time campaign. Or let's say I need to get all my customer names in sync, but then new data quality features in my applications will keep those records clean. I can do a one-time data-cleansing project on a cloud platform.

Why can't you see big companies turning to the cloud as their primary platform?

Large companies with big data centers can probably do it as cheaply or cheaper than a cloud offering. If I had to sum it up, I'd say there is a future for the cloud, but initially it's going to be mostly vendors using it for developing third-party software and doing proof-of-concept projects. It will then move into development projects for corporations. Finally, it will move into short-term projects and platform use by smaller companies, but it will be a long time before we see it go much beyond that.