Considering how much I like the sound of my own voice, it may surprise readers to know that this is my first technology blog. It didn’t take long for me to decide that some thoughts on Software Defined Data Centre and the changing infrastructure landscape would be the subject of the first entry; after all, it’s Ethos Technology’s raison d'être (four lines in and already I’m sounding pretentious?!).
I’m continually surprised by how often tech vendors and analysts get caught up in debating what does or does not constitute ‘true’ software defined, whether it be networking, storage or something else; in my opinion this misses the point of what we’re trying to achieve with the next stage of data centre evolution. At the heart of the matter is a desire to move away from the stitching together of discrete, monolithic systems to distributed architectures; from imperative to declarative models of policy definition. There will always be a combined hardware and software element, where the line between them is drawn is largely arbitrary; the key is that a solution abstracts policy and requirement from the complexities of the underlying technology.
It was appreciating this and developing an understanding of some of the nuts and bolts of products I’ve designed solutions around which lead to a bit of a watershed moment for me. I’ve spent the bulk of my career as an infrastructure architect; for much of it, I perceived specialisms such as application development as requiring a completely separate set of skills and understanding. I now have a different perspective on things; when I looked under the covers of technology I was working with, much looked familiar.
The mathematics underpinning the Paxos algorithm is the same whether it’s used to resolve consensus in a database, a storage cluster or an SDN control plane. Technologies like Cassandra serve as a foundation for use cases as diverse as commercial hyper-converged solutions to the data analysis engines used in the CERN large Hadron collider. High availability and consistency are key whether at the database, storage or network tier and often the techniques used to provide them aren’t a million miles away from each other, irrespective of the layer in which they’re used.
I look at the melting pot that exists now and I’m reminded of an idiom a friend of mine in sales (I don’t hold that against him) often repeats, “Change creates opportunity”. Now, whilst I do agree, he should probably add that opportunities usually offer as much potential to succeed as they do to screw up royally. I’m sure that both roads will be well trodden over next few years; fingers crossed Ethos’ first independently developed platform spends more time walking the first (watch this space...!).
Wherever you look the lines between disciplines are becoming blurred. The DevOps mantra of ‘Infrastructure as Code’ is both catchy and prescient. As the DC evolves the techies in our industry must evolve and diversify too. The virtualisation guy had better know his VXLAN from his NVGRE; the network guy Python from Ruby; the developer, VM from container; if they don’t they are going to be left behind. The kids coming up have been immersed in technology since before they could walk, in today’s world even D.A.R.Y.L would be second string...