IT has once again found itself engaging in the age-old practice of "white boxing." This time, however, we’re not piecing together our servers; virtualization has instead driven us to white box our entire data center infrastructure.
The end result of white boxing servers often ended up being a data center full of inconsistencies, with additional administrative overhead and an increased chance of failure following every configuration change. This is why converged infrastructure has quickly become our industry's hottest topic, as white boxes are now being assembled to create virtual environments.
A few years ago, we discovered all the wonder and excitement associated with virtualization. In turn, virtualization was implemented as soon as possible on any available hardware. But the piecewise addition of more SAN equipment and networking, along with each new round of servers, created more and more colossal interconnections.
Before long, the complicated web of do-it-yourself virtual hardware stopped being wonderful and started being problematic. Doing it yourself, as we’ve only now rediscovered, doesn’t scale .
Solving this problem and beating back our industry’s second generation of white boxing is what converged infrastructure intends to achieve. That’s why, rather than "converged infrastructure," I prefer the term "hardware that's designed with virtualization and cloud in mind."
Converged infrastructure in a nutshell
At the end of the day, your virtual infrastructure or private cloud runs on monitoring data. That data explains how much capacity is on-hand, broken down into major categories: Compute, memory, networking and storage. Call this your supply of resources.
It also knows how many of those resources your virtual machines (VMs) require. This second number represents demand. With supply and demand now abstracted into some set of numerical values, you have now generated an easy-to-understand "economics of resources" that represents the state of your data center.
In a converged infrastructure, this provides a recognizable warning as to when more resources are needed. Trending the use of resources means knowing that more networking, storage or compute power will be required at a specific time. Supplied and consumed resources are now quantified and used, rather than organizations relying on best guesses. Trending also makes purchases substantially easier to plan and budget.
To achieve its goals, converged infrastructure hardware is completely modularized, not unlike a stack of Legos or Tinkertoys. Each module connects with minimal effort into the greater whole that is your data center, much as an additional hard drive is snapped into a server or SAN today. The modules also contribute a known quantity of resources, increasing your economic supply. You can add to your total storage, computing power or memory in the same way.
More importantly, each module is something you'll purchase by popping over to a manufacturer’s website and clicking "buy." You’ve done this for years with servers; why couldn’t you do it with your entire data center? What will arrive are ready-to-insert components with minimal cabling and trivial installation. Wrapping around this entire system is a management toolset that recognizes new hardware and seamlessly adds it to your pool of resources.
And this isn't all in the future. For some manufacturers, the hardware is already here. For others, it's on the roadmap. Some of the components -- blades, modularized storage, dense networking and so on -- are being advertised by the major manufacturers, even if they haven’t yet explained how this new approach works. And the management toolsets are also well on their way.
With names like BladeSystem Matrix and Advanced Infrastructure Manager, these prepackaged virtual computing environments manifest the resource economics at the hardware layer while your hypervisor management tools deal with individual VM actions. The combination of these two pieces is the source of what we now think of as private cloud computing. Converged infrastructure is just the enabler.
So is converged infrastructure a fancy name or an actual technology? In a way, it’s a bit of both: not there to eliminate hypervisor management but to augment it. Converged infrastructure’s hardware and management tools provide a way to end, for the second time, our nasty practice of white boxing.
About the author:
Greg Shields, Microsoft MVP, is a partner at Concentrated Technology. Get more of Greg's Jack-of-all-Trades tips and tricks at www.ConcentratedTech.com.
Converged infrastructure can be the cure for complex IT environments
IT architects ponder converged infrastructure "pods"
Converged management, not infrastructure, holds appeal
How a virtualization stack provides simplicity, scalability for VDI