Software as a Service databases wil also have a impact on DBAs.
I call these the “Google Databases”, picking on the most notorious supplier of SaaS.
Part of what makes SaaS databases attractive is the cost savings through automation. In a recent podcast, Jonathon Lewis talks about clients with 500+ databases and only a handful of DBAs. If those 500+ databases support 25 SaaS customers where there would have been 25 or 50 or even 75 DBAs, we can see a big impact.
This is also a side effect of grid computing. If database suppliers like Oracle make it easy to consolidate and manage large numbers of databases in grids, there has to be a way for smaller customers to consolidate their databases on those grids. A lot of customers don’t have hundreds of databases to take advantage of this economy of scale. Buying services from somebody else with hundreds of databases lets them tap into those savings.
This isn’t new a new idea really. Back in the mainframe days many customers could not afford or use an entre big IBM mainframe. The timesharing business sprang up to sell time on one mainframe to many customers. Sometimes companies sold their unused capacity, other times timesharing suppliers bought a mainframe for the purpose of sharing it out. Now we see the same effect with large grids shared out as SaaS.