Contributed Article by Daniel Heimlich, Vice President, Netuitive Inc.

The Virtualization "Tipping Point" in 2010 

At ITXPO in Orlando this year, Gartner’s Tom Bittman stated that most enterprises have only virtualized about 16% of their x86 workload (see it here on YouTube).  Gartner also expect that number to reach 50% by 2012.

But many enterprises are going to hit the wall at about the 30% mark in 2010 because they will hit a “tipping point” where the complexity of the infrastructure will outpace their ability to monitor the environment using traditional technologies.

Here’s the story of how one company hit and got past the roadblock at 30%, as told to me from the viewpoint of their VP of Virtualization Operations. The company is a large financial institution with over 50,000 servers deployed globally.  After some initial success and tremendous ROI, the support for virtualization now goes all the way to C-level Execs.

At the beginning, the Virtualization Team did what most people do: they targeted low-hanging fruit in the form of low-utilization servers in non-critical roles (print and file servers, etc).  Then they began to go after some applications with the goal of one day targeting even the most critical ones.  Here’s where the "Tipping Point" was reached.

At about the 30% virtualization mark (based on total # of servers), the team hit the wall. At this "Tipping Point", when performance issues hit, it became difficult or nearly impossible to diagnose problems with conventional monitoring tools (with dashboards, dials and fixed performance metric thresholds).  Is the VM problem the problem or the host? Or is there a problem in storage?  Or in the network?  Is the problem on one VM related to another?  Can a series of minor recurring problems mask a looming major one?

As their VP of Virtualization Ops stated: 

"We were experiencing what others had warned me about. We suddenly realized ‘wow, this is an extremely complicated environment’ and when there’s a problem, you get 8 staff and 3 vendors on the phone all pointing fingers at one another.  The need for cross-silo insight became absolutely critical”.

The team went looking for a tool that could provide automated monitoring intelligence and analytics to help them monitor performance in the environment.  Without such a tool, they felt that they were not going to get past the 30% virtualization mark without doubling their staff – and that was not going to happen.

After a thorough investigation of 5 different solutions, the company selected Netuitive’s self-learning performance management software – the “Best of VMworld” winner in 2009 for Virtualization Management (www.netuitive.com).  Netuitive is used to monitor performance across the entire virtual data center, including server, network and storage infrastructure.

Netuitive helped the Virtualization team win over skeptics among the application owners.  "These [Line of Business] guys have their own flesh-eating IT.  They want to see data and multiple layers of the infrastructure; they want to know you can respond in seconds, be proactive, and stay ahead of the curve."

So to reiterate my prediction for virtualization in 2010:

The good: many more enterprises will reach the 30% virtualization mark for x86 server workloads (up from 16%) in 2009.

The bad: they will hit a roadblock due to reaching a complexity "tipping point" in the infrastructure

I think we are going to have a great year at Netuitive.

About the Author

Daniel Heimlich is vice president at Netuitive, which is the leading provider of self-learning performance management software. He has 15 years experience with infrastructure technologies and IT performance management. Prior to Netuitive, Daniel was at Citrix.