I was privileged to deliver a session entitled Managing SQL Server in a Virtual World at the PASS Summit 2012, the largest annual conference for Microsoft SQL Server. It was a packed house, literally at standing-room-only capacity. I delivered the session with my friend David Klee (http://Kleegeek.com) and we were swarmed by attendees after the session wrapped up.
With almost 600 people in the room, we conducted one of those informal polls that speakers like to do along the lines of “Raise your hands if …" and the informal findings were very telling. Probably around 90% of the attendees used VMware and SQL Server in some capacity and at least 60% used it in production environments. Another important fact was that only 10% of the attendees were actually able to get information on the performance of the actual VMs themselves. Most had to get all of their information and support from the VM / System administration staff.
So while it was clear that the last bastion of the physical machine, database platforms, is finally ceding ground to the virtual machine, it was also clear that DBAs are being denied the tools to make them successful. Here’s the good news - VMware is aware of the gap in tooling for DBAs and help is on the way!
On Tuesday, November 6, VMware announced via blog post (http://bit.ly/W5VU9y) that they have released vFabric Data Director 2.5 with support for SQLServer and Hadoop. vFabric Data Director 2.5 alleviates database sprawl and inefficiency, providing enterprises a fast, simple and cost effective database-as-a-service that accelerates application development. Now, this newest release extends the benefits of virtualization to traditional relational databases like Oracle, Microsoft SQL Server, and Postgres, as well as the popular noSQL platform Hadoop.
I attended an extensive demo presented by Gopal Ashok, senior product manager for vFabric Data Director, and was very impressed with its features. Significant among these features was the ability do a great many activities, such as spin up new, cloning, or recovering existing SQL Servers nearly instantly, using standardized templates.
VMware’s Fausto Ibarra, senior director of product management, shared his thoughts with me on the growth of database-as-a-service and its implications for the enterprise:
"Database management is an area that's ripe for disruption. Enterprises need to be able to quickly and cost effectively provide a policy-based database as a service for most common data technologies to their developers from a single web portal. Data Director enables enterprises to do just that - easily deploy and scale Hadoop clusters on their existing VMware vSphere clusters, reducing the time to insight and eliminating the need for dedicated physical infrastructure.”
The attendees at my PASS Summit session would heartily agree with that sentiment and would even suggest that this trend is accelerating. I’ve posted my slides at http://bit.ly/W5TDex if you’d like to see them.
So what are you seeing? Is virtualization adoption increasing for your database applications? If not, what are the impediments that are preventing adoption? I look forward to hearing back!