I love technology, I love this industry, and I’ve really enjoyed the movement I’ve seen particularly since the advent of virtualization. The development, the movement into less tinker-toy, and more defined architecture has made the things I used to do on a regular basis far less complex and more to the realm of plug and play. Of course, we’re not entirely there, but we are moving in that direction.

Along with the revolutions in the architectures (storage, compute, and networking) has come a lot in the way of “also-runs, and wanna-be’s.” It’s an industry with more jargon and buzzwords than any I’ve experienced. One of my biggest frustrations is not that companies assume they can build something better, (that’s really what competition is all about) but that so many of these buzzwords which get assigned become overused, and inappropriately absconded.

Software Defined originally had some meaning. In Networking, ACI, OpenFlow, and NSX are actually software defined. Of course, at some point, they rely on hardware, but the idea that the removal of the control plane from the silicon is what Software Defined really means. Storage gets a bit more convoluted. When I worked for Nexenta, I felt that the application of that Software Defined Storage moniker actually had meaning. The model included commodity off the shelf equipment. Simply put, some compute power, cabled up to a JBOD. All the magic happened in the software. The deployment of block or NFS, the provisioning of that toward its intended target, even the policy of how much IO to push toward the volumes all took place in the software. To me, the model, while a bit complex, made sense. But, of course, we had a multitude of competitors who saw the elegance in Software Defined Storage, and packed their bags toward the SDS bandwagon. Some legitimately, and others not even remotely accurate. It became almost a religious war on who is, and who isn’t promoting SDS with integrity.

Well, along came the Nutanix and Simplivity model. It’s called Hyper Converged Infrastructure. HCI is a model that in a single building block includes the Compute, Storage, and Networking all in one pre-sized prebuilt building block. It made and still makes sense. All sorts of cool tools can be built into these, whether software or silicon based. Tools like replication, deduplication, etc. have value and make for the features, advantages, and benefits of one package over another. This is the true competitive landscape and the decision-making process for the customer. I appreciate all of that. But the landscape is filled now with quite a number of fantastic architectures that really are not HCI. For example, there are storage platforms calling themselves hyper-converged, wherein the elements of compute and storage are separate and discrete components. How could this be hyper-converged? Truthfully, it isn’t. While, many of these compelling architectures are so very workable, and solve so many problems, that really doesn’t mean that they can climb on the Hyper-Converged bandwagon, does it?

OK, who am I to say? I know, I’m just a guy who’s got a view of the landscape. But I have always felt that words mean something. If just any architecture can glom onto a name and claim legitimacy, even if the architecture doesn’t have a legitimate claim, then the muddying of the name further means nothing.

Again, I’m not actively disparaging any particular build. All I’m really saying is that sometimes an attempt at publicity or attention, simply to be noticed is not good marketing.

I wonder if my words will mean anything to anyone.

I wonder if my words will mean anything to anyone.