Database

SQL Server Consolidation, Part 2

December 20, 2012

SQL Server Consolidation, Part 2

Consider for a moment that you have a deck attached to your house. It is one story above ground level and is growing weaker with each passing year. The former homeowner painted (!) the deck, making your options for upkeep quite limited. And what would those options be? Well, most people would look to have their deck updated. Of course, they could also consider having the deck removed altogether. Another option would be to have the deck reconfigured while it is being updated, perhaps even have a smaller deck built.

Now, examine the deck closely. You will find that while you speak of the deck as one complete unit, it is actually made up of several distinct pieces of wood. Some comprise the floor, some the railings, and some are stairs. If you go one step further, you can consider your deck as a data center and the pieces of wood as individual servers. Yeah, it’s that easy.

So, you have all these servers that make up your data center. Some servers are better than others, some are old and weathered and need to be replaced as soon as possible, and some are slightly newer. So, what do you do with your deck? More importantly, what are you going to do with your servers? What are the most important decision factors?

Costs Factors

No surprise here, right? Not only is cost one of the motivating factors as described in part one of this series, it is also the biggest decision factor. You will need to consider the costs associated with everything. While the idea of logically consolidating three servers into one looks great on paper, there are lots of hidden costs that people fail to consider.

First up is people’s time. It takes time to consolidate systems; they don’t just migrate themselves. Next up is new hardware. Chances are you don’t have the right hardware in place already to serve as a virtual host, so you are going to need to make some initial purchases. And then more once you realize that you want little things like disaster recovery and/or high availability. And while you are adding in all of those costs, you need to consider the cost avoidance of things. For example, if you know it will take a year to consolidate three servers into one and cost $500k to get the job done, but the end result will be a cost avoidance of only $100k over the next three years, you will want to reconsider your project. However, if the end result means less overall support, increased up-time, and a more stable environment, then you may consider it to be $400k well spent in light of other options. Going back to the deck, do you go with ipe, cedar, teak, or a composite decking material? Each has a benefit over the others, and each has a distinct price. Don’t forget the railings, along with the yearly maintenance: all those costs add up over time.

Stability

We are all adults here, right? We want our environments to be stable. We want to limit access to servers and systems, only letting the few necessary hands touch things. If you have server sprawl, you have way too many operating systems to maintain, making patching more difficult, not to mention security. With consolidation, you get back some of those operating systems. With virtualization, you get back your time. In exchange, your users get back a more stable environment. What’s that? Some business-owned application just loaded 500GB of files and filled up your drive? No problem, we can expand the drive quickly, even change that RAID5 to a RAID10 while the server is running. You say a developer claims the box is out of memory? No worries, I can allocate more memory to that guest right now, without leaving my desk. Did you say the hardware lease on that host has expired? Well, let’s just migrate those guests off that host and onto a newer one without ever interrupting the existing connections.

Ease of Implementation

What tools will you be using if you start building virtual hosts and guests? How will you monitor this environment? What virtualization software will you be using? Will it scale enough for your needs? Most importantly, what can we implement today, given the current configuration of our shop? It is similar to asking yourself if you know what you want versus what you can afford. Sure, we all want the most expensive hardware possible, but what would it matter if you had no place to put it, or the right people to run it?

Now that we have discussed some of the drivers for consolidation projects, as well as some of the major decisions factors, the last entry will discuss the overall process that can help you to efficiently evaluate your current systems.


Thomas LaRock is a Head Geek at SolarWinds and a Microsoft® Certified Master, Microsoft Data Platform MVP, VMware® vExpert, and former Microsoft Certified Trainer. He has over 20 years’ experience in the IT industry as a programmer, developer, analyst, and database administrator. LaRock has spent much of his career focused on data and database administration, which led to his being chosen as a Technical Evangelist for Confio Software in 2010, where his research and experience helped to create the initial versions of the software now known as SolarWinds® Database Performance Analyzer. LaRock has served on the board of directors for the Professional Association for SQL Server® (PASS), and is an avid blogger, author, and technical reviewer for numerous books about SQL Server management. He now focuses on working with customers to help resolve problems and answer questions regarding database performance tuning and virtualization for SQL Server, Oracle®, MySQL®, SAP®, and DB2®. He’s made it his mission to give IT and data professionals longer weekends.