Hyperconvergence (or Back to the Future, Part Deux)

NO3N2Getting in on the latest “thing” is always interesting, especially when you take some time and look back at old YouTube videos of presentations at various trade shows. That is when you can really see what was reality, and what was not. In the case of hyperconvergence we see that the idea of going back to the commodity platform and building a webscale infrastructure was not only dead-on, but it is drastically picking up steam at a break-neck pace in many environments that were formerly very traditional in their approach.When I joined EMC‘s ScaleIO team, the idea of hyperconvergence was not really new, but it was new to the Enterprise. It is something that had gone on at many of the cutting edge tech companies in Silicon Valley for some time in one way or another. Maybe not in the purest form of hyperconvergence, but certainly the idea of leveraging low-cost commodity hardware with software on top of it to facilitate all of the things that we came to know and love about Storage Area Networks.

Virtualization was a big step in this direction, and the advent of OpenStack on VMware was huge, but the idea of Software Defined Everything is game-changing. The idea that I can have ANY hardware platform I want with a software layer on top of it that virtualizes my server infrastructure…that’s a big deal. That means my compute engine is just that…an engine, and as I need more power or more storage, I add the nodes and spindles as I need to. My engine just keeps getting bigger to handle the workload I need to handle.

I think we have all said it a million times. It’s back to the Mainframe all over again. One big hardware platform running a layer to virtualize, then breaking up the resources as needed. Instead this time our big hardware platform is made up of a bunch of smaller systems acting as one. In this model with a VM layer on top of it we have our Virtual Server Infrastructure (VSI), and frankly, there are very few (and I mean VERY few) things left that cannot be virtualized.

Years ago I was at a VMware User Group meeting in Orlando when I met Stephen Beaver who was at the time with Florida Hospital. Stephen is now the CEO of BeaverDam Consulting in Orlando, FL. He was recognized early on in the virtualization world as one of the best, and with his operational and technical knowledge being well know in the virtualization community, his statement “If you can’t virtualize this application, why?” began to get traction all over the place. I can’t tell you how often I would hear Stephen and that statement quoted all over the Southeast. It was that statement that he used to transform the way Florida Hospital ran their IT infrastructure, and it really was prophetic, because when you think about it, that is what we are trying to tell people today. Virtualize EVERYTHING…and then move to the Cloud (though a lot of times vendors leave that last part off).

Virtualize EVERYTHING. Not long ago I would have stood up and gone toe to toe with anyone who said you can and should virtualize Hadoop, SAP HANA…anything like that. Now, I find myself repeating the statement from Stephen…If you can’t virtualize this application, why? Simply put, hyperconvergence has allowed us to get better efficiencies, faster speeds (network and storage) for less money and a more simplified operational model. Virtualization was just the beginning, but it is the very foundation for everything that we do today. Hyperconvergence does not require virtualization because you can certainly enjoy the benefits of hyperconvergence on bare metal servers. Why you would not take FULL advantage of hyperconvergence in a virtualized environment is the question.

Maybe it is time to ask yourself some questions when it comes to your IT environment like “if I am not fully virtualized, why not? Is there a valid technical reason that I cannot be?” If you are fully virtualized, are you going to take the next logical step of hyperconvergence, and removing the restrictions that traditional Storage Area Networks and hardware place on your IT environment, or are you going to take that next step toward webscale?

There is a reason companies like Facebook, Google, Yahoo and eBay consistently get the attention of everyone running a datacenter. It’s because they realized that there is a better way to have IT work for them. Hyperconvergence is that way, and it can work for you too.

1 thought on “Hyperconvergence (or Back to the Future, Part Deux)”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s