For the record, I’ve never heard of the British band “The Status Quo“. Apparently they are a pretty big deal in the UK, but unlike say the British Invasion bands (The Who, Beatles, The Rolling Stones, or even later Pink Floyd, ) they never really caught on in the states as far as I can tell. So when I went to search for some imagery around the “Status Quo” this is what I got.
To be honest, the idea behind this blog post has been kicking around in my head for the better part of three months now since I joined SimpliVity, and I’m just now getting around to putting it down in words. The last 3 months have been nothing short of crazy. As anyone who reads this blog or has followed me on Twitter knows, I’ve been doing a bit of travel. I’ve been speaking at VMUGs in the western states, I’ve continued down the vExpert path, and I’ve been trying to capture all of this when I can here on this site.
I’ve purposely tried not to speak too much about the technology I’m currently working with. I’ve done this not out of disinterest or lack of desire to do so, but primarily because I think that it’s important for me to keep a level of impartiality in my personal space. So while I may write at times about my adventures in the Sales Engineering/Technical Marketing space, I’ve tried not to pimp my ride as the kids these days say. With that said, today I’m going to do a little bit of pimping.
Converged Infrastructure vs Hyper-Convergence
I’ve been giving presentations across the Western States as of late primarily around the topic of Converged Infrastructure and Hyper-Convergence. The latter term is one that was new to me prior to taking on my current role. While I had heard of several of the companies that are working in this space, I’d never heard the term “Hyper-Converged” before. If we look at TechTarget (seriously these guys need to lay off the spam parade when you sign up with them) we get this generalized definition of the “Hyper-Converged” spectrum:
Hyper-convergence was born out of the converged infrastructure concept of products that include storage, compute and networking in one box. Systems that fall under the hyper-convergence category also have the hypervisor built in and are used specifically in virtual environments. Storage and compute are typically managed separately in a virtual environment, but hyper-convergence provides simplification by allowing everything to be managed through one plug-in.
This sounds about right to me, but lets take it a few steps further. While the definition above gives us a general understanding about what the basic concepts within a hyper-converged infrastructure are, it doesn’t quite go far enough for my taste.
If we start to look at the landscape today in many IT organization, we see a common thread, and that thread tends to be that we have taken disparate hardware from several vendors, or a master vendor, and mashed it together to serve the many needs of the datacenter, ie: Storage, Compute, Networking, and either a Hypervisor, or some form of OS pre-integration. This is typically the “reference architecture” approach to Information Technology, and its served us well for the last decade plus. To me, it represents the “Status Quo” for many organizations. I need storage, I buy storage. I need servers, I buy servers, I need networking I buy networking. And so on and so forth. It’s a model that has made many a company many a dollar, and for the most part it works pretty well. But…
And there is always a but….
The model is broken and today we are looking to have greater efficiency within the datacenter. Converged Infrastructure provides assistance in the simplification and the breaking down of the complexities within the standard Reference Architecture model. If I look at the convergence landscape this is what I tend to see: (thanks to Stevie Chambers for his piece that helped define the below points)
- Semi Converged: product is not complete when received, has limited capabilities and capacity. Significant work is still required to deploy workloads.
- Fully Converged: has standard network, compute, storage, and virtualization capabilities when received. Still may need additional configuration to deploy workloads.
- Super Converged: has additional feature sets beyond Fully Converged such as data protection and should be nearly turn key to deploy. Plug and play for the most part.
- Hyper Converged: all the benefits of fully converged infrastructure but based on small form factor appliance providing datacenter in a box functionality. Rack, and run deployment.
When we start to look at solutions that fit into these definitions, the generalized landscape should look something like this:
Now the definitions start to make a little more sense, especially when we allocate for their deployment scenarios as well as their capabilities. The Degree of Efficiency, coupled with the Scope of Functionality are the two factors to take into account. (WordPress formatting sucks so you get a screenshot)
Looking at the graph, the Degree of Efficiency and Scope of Functionality should make more sense now. Specifically we start to see how the primary convergence definitions find their place on the axis.
Today there are many solutions being offered by many vendors that address specific points on the graph, Moving from your standard reference architecture that may touch on some converged aspects, out to Hyper-Convergence. Even still, within the spectrum of the Hyper-Convergence realm, not all solutions are the same, and not all solutions will touch on the various points on the Scope of Functionality axis.
Moving Beyond The Status Quo
Much of what I’ve defined above is well outside the “Status Quo” of the technology solutions we have been building over the last decade. Simply put, as humans we have preconceived notions and biases about the different vendors and technologies that exist today. The commonality of EMC, Netapp, Cisco, HP “shops” is something that I would consider to be commonplace within many organizations. Not to say there are not good reasons for this line of thinking. There is the benefit of a platform that has been qualified within the organization. Familiarity with the management tools, Histories with support groups. Etc. etc. etc. These factors weigh heavily in modern infrastructure, primarily because we have been taught that the underlying hardware platform has actual differentiation and benefit. This is the Status Quo, and its the thing I fight against daily.
This is why I find the aspects of a Hyper-Converged infrastructure so compelling. If I can take my traditional stack, collapse it, and deploy the same functionality based simply on need (compute/storage) then that becomes a very powerful tool at my disposal. Today I have that powerful tool in the form of the SimpliVity OmniCube.
It is not a simple task to change peoples minds when they have been utilizing a model that has worked for them for years. The battle I face moves well beyond the basic FUD that gets thrown about, the battle I face involves winning the “hearts and minds” of the IT staff, and pointing to a different way to manage, deploy, and enable workloads in their organization.
So with all of this said, this is the first in several posts that I will be doing that addresses why I think the old way of thinking is archaic, inefficient, and simply not going to benefit you or your organization in the future. The move towards simple deployed resources needs to be addressed. We need to start looking beyond our historically closed deployment models, and start to address what I see as the actual need, facilitating the IT shop to provide resources that meet the performance requirements of the workloads within the organization without unnecessary added complexity, clutter, sprawl, and inefficiency.