One of the trends that I see that is going to specifically change the way database administrators work in the future, is these new storage arrays. I’m a real fan of Pure Storage so I’ll use them as an example. Pure is an SSD based storage that’s able to reduce your costs and more importantly improve performance. I worked with VMware to get together a room full of SQL Server MVPs to look under the covers of their technology. Pure was kind enough to provide the storage array, and as we got this room of about 15 of us together and we started to try to put a workload on the storage array, we couldn’t even bring it to its knees. Of course given enough time, I would argue that any technologist could bring any technology to its knees – but what really intrigued and impressed us was that the more we threw at the storage array in terms of IOPS, the more this thing thrived. We never saw the stress point.
The most common source of problems in a virtualized environment is the storage array. Another common source of problems is the IO. As a DBA, I can tune the IO or I can tune the CPU but I’m always making trade-offs (for example adding index to minimize IO). Now as I look at these storage arrays and how they’re evolving, it’s clear to me that in the next 5 years DBAs aren’t even going to think about the storage layer. It’s just going to perform at a level that’s appropriate to the database and I’m no longer going to have to go do things like separate logs from files. If I need 50,000 IOPS, I won’t worry about where it gets stored or if the storage array can support it.
Now take it to the next level. You now have hyper-converged architectures where you have all the layers of a traditional, multi-stage environment from CPU, memory, etc. in one simple device that’s easy to manage. Nutanix and SimpliVity are good examples of this.
Really interesting things are happening on the technology front, specifically around storage arrays and hyper-converged architectures that can be game changers for how database administrators work and how we manage databases in the future.