Why the heck does infrastructure still matter?

Adam CatbaganWritten by | Uncategorized

There is no denying that the IT industry is undergoing a massive change. You could make a pretty convincing argument that the notion of a software defined data center is the heart fueling this monumental shift. This is filling the minds of business leaders around the globe with thoughts and ideas of consuming IT as service, making the ever so elusive goals of reducing risk and cost while increasing IT agility an achievable reality for IT business unit leaders. But as an IT thought leader who has built a career inside the data center, I find myself asking this question: Does infrastructure still matter?

Well, you can bet your bottom dollar that it does! In fact, I’d submit that infrastructure matters more now than it ever has before. For a SDDC platform to be successful, there are two key elements that are essential: First, the business must have agile business models in place. Second, the business must have agile infrastructure. This will include plans for media transitions (think moving from HDDs, to SSDs, to NVMe intelligent drives); inclusion of both scale-up and scale-out architectures to meet workload demands; and management and orchestration tools to deliver IT services back to the business.

Compute, Network and Storage

I believe the three main elements of a SDDC platform are compute, network and storage. Most of us are comfortable with the idea of virtualization, the process of creating a software-based (or virtual) representation of something rather than a physical one to boost efficiency and agility. This same idea holds true for software defined network and software defined storage constructs, which leads me back to why infrastructure still matters.

Would you ever buy a car without a gas pedal? You may have the best engine, the best break system, the best tires… but if you’re missing the thing that makes the car “go,” you might as well just push the thing off a cliff. As such, you might be partnered with the some of the best SDDC platform companies in the market. But, if your infrastructure is missing the hooks to make your SDDC platform “go,” well maybe, just maybe, you’d better get to pushing that infrastructure off the cliff right along with that car.

Adam Catbagan

The White Box Infrastructure Trap

It’s been said that in a typical IT organization, 70 percent of every dollar is used just to keep the lights on. So, every CIO is looking for ways to drive down costs from their budget. One of those ways may be to use a white box compute manufacture. After all, compute is compute is compute, right? Not to mention the allure of lowering compute costs by selecting a cheaper compute manufacturer! Well, that just might earn you a ticket to this year’s employee recognition party.

As we say here in the Colorado mountains, “Truckers don’t be fooled, steep grades are still ahead.” Or, in other words, when considering utilizing white box infrastructure with an SDDC platform, you must consider the following:

  1. How will you handle hardware patches
  2. What type of power management is available for this hardware? Is this per node or is there a centralized repository all nodes will report into?
  3. What about system management?
  4. Has this infrastructure been certified and validated to work with my chosen SDDC platform software for orchestration and automation?

Oftentimes, what seems to be the easy button ends up costing the organization more in capital expenditure and employee morale.

The Buy vs. Build Conundrum

Once you’ve chosen your SDDC platform, solution deployment presents two options: You either build it or you buy it. Let’s go back to that car analogy, but this time the car has all the parts. If you’re a super mechanic and have the time, you could order all the parts, have them delivered to garage, then begin assembling your vehicle bolt by pain-staking bolt. Or, you could stroll on down to your local car dealer, purchase the car that’s right for you and drive it home that very same day. Oh, and did I mention that the new car will come with a manufacturer warranty in the event anything goes wrong? Neither way is wrong, but one must weigh the pros and cons of both options.

Coming back to the SDDC platform options, there is a do-it-yourself approach. By combining discrete solutions available from the line cards of various hardware and software manufacturers, one can build an infrastructure stack that offers the business choice and avoids a potential concern of vendor lock in.
There are also solutions available today from leading converged and hyperconverged manufacturers that greatly simplify and combine the SDDC components of compute, network, storage and orchestrating in a single “box.” These solutions are preconfigured, prevalidated and come with single entity support.

Essentially, you are buying the business assurance that these platforms will work and be supported as sold.

Where Should I Start?

The road to SDDC nirvana is riddled with obstacles both large and small. It’s not for the faint of heart, and, at times, one may ask, “Did I make the right choice – am I really doing what’s right for the business?” Rest assured, you’re not alone. Thankfully, our friends at Intel have provided a checklist of sorts that may help you on start your journey to SDDC relevance:

Arrow can help you navigate the road to a software defined data center. Please contact your Arrow representative with any questions about SDDC.

By Adam Catbagan
Sr. Systems Engineering Manager
Arrow ECS

Last modified: May 3, 2019