As I walked through my data center a few months ago, I reminisced about how small it was just a few years ago and winced at the huge expenditures the organization had made to grow it to its current state. What made matters worse, I was making the walk-through to plan yet another expansion and winced yet again when I thought about how I was going to present this requirement to my budget conscious CIO.
It didn’t take me long to decide it was time to think strategically and make server consolidation and virtualization a project under its own merit and funded as a strategic corporate initiative. This would remove any objections from individual project owners about the few extra pennies that it might cost to implement their pet project, because in the great scheme of things, it would reduce overall IT department costs and subsequently benefit the corporation as well.
Costly data center build-outs and space-devouring server expansions are a thing of the past. Jerry Hodgen shares his tale of planning a server consolidation project and getting everyone, from prickly vendors to penny-pinching CIOs, onboard.
“I called my software vendors together and invited them to join us or get out of the way because I intended to steam forward with or without them.”
I won’t bore you with a long-winded explanation of the difference between server consolidation and virtualization, but I will give you my simple definition. To me, server consolidation is to condense a number of servers in a smaller amount of space. In today’s technology that translates into blade servers, which Dell, HP, and IBM all offer. Server virtualization involves emulating multiple servers on one hardware platform. In this market, there are a few vendors including Microsoft and VMware. However, in my opinion, VMware wrote the book in this arena and is leading the charge.
In my mind, server consolidation and or virtualization seemed to be a much better approach than expanding data center real estate. My team had dabbled in blade servers and virtual machines over the course of the past couple of years, so the concept was not really new to us. As a matter of fact, we had even deployed a few blade servers and a half dozen or so VM Boxes, but we had not really taken a serious look at the technology before now for a couple of reasons.
First and foremost, many application providers had gone on record in the past about not supporting their applications on virtual machines. Secondly, in the case of blade servers, the high cost of entry for the initial blade server enclosure made it difficult to justify on a project-by-project basis.
Take Charge of the Situation
The first thing I decided to do was to take emotion out of the decision by engaging a third-party consultant to assess the server infrastructure. This would free up my team to continue to take care of day-to-day business while highly qualified engineers focused on the challenge.
In addition, I called my software vendors together and had a serious discussion with them concerning my intentions and invited them to join us or get out of the way because I intended to steam forward with or without them. To make a long story short, all of the vendors eagerly jumped onboard and discounted any Doubting Thomas attitudes that some of their support staffs may have displayed in the past.
It didn’t take long for the consulting company to deliver the results of the study I had commissioned. Frankly, I was pleasantly surprised with the results — 170 servers of 195 servers were candidates to be virtualized. Servers that didn’t make the cut were generally systems that housed high transaction and/or CPU intensive applications. After reviewing the preliminary results, my team and I decided to capitalize on our existing infrastructure and experience.
As a result, we settled on HP blade servers as our server standard. This was an easy decision to arrive at because we already had a traditional HP server environment as well as HP’s Insight manager system. A key factor in making this decision was HP’s wise decision to include Cisco routing technology in its platform, which allowed the blade servers integrate into our Cisco LAN and WAN infrastructure seamlessly.
Leverage Existing Infrastructure and Expertise
Now that we had decided upon the Blade Server strategy, we had to take a hard look at what virtualization platform to use. There are a number of players in this space including Microsoft. In the final analysis, we felt more comfortable with VMware not only because of our past experience, but VMware’s capability to perform equally well with both Microsoft and Linux.
“I could tell I commanded the CIO’s attention by the scowl on his face and the way he shook his head, as if that would make these issues go away.”
In addition, VMware is at the forefront in promoting and advocating standardization for APIs related to VM machines. A vendor striving for standardization and interoperability goes along way in my book.
A Fiscally Responsible Business Case
Now that I had all of my facts together, it was easy for me to put together my business case to the CIO. Bear in mind “business case” to my CIO translates to “accomplishing more while reducing costs.” I started off my conversation by discussing the number of applications we had to deploy in the last year and reminding the CIO about the new application requirements that business operations discussed with us in last month’s business plan meeting.
I followed with the dire fact that our data center had reached capacity. I could tell I commanded the CIO’s attention by the scowl on his face and the way he shook his head, as if that would make these issues go away.
Then I presented him with my team’s plan to embrace a server consolidation and virtualization strategy, which would shrink our data center space by more than 50 percent and reduce annual operating costs by more than $400,000 per year. Needless to say, the scowl immediately changed to a smile.
And I swear I could see, in the glimmer of his eye, visions of the CEO presenting him with a big fat bonus check.
Article appeared originally on Enterprise IT Planet.com.