The Walking Dead is one of my favorite television shows, and is a perfect metaphor for the IT industry. We’re at the beginning of one of the major shifts in IT spend, which only happens every twenty or so years. It is the migration from scale up and limited scale out virtualized environments, to hyperscale or web scale computing. The consumption of applications at global scale with millions of interactions and the move to post-pc devices breaks the client server model of computing. Since 1997 I’ve been working at web scale, when I literally had a crawl of the entire Internet in the basement of Alexa. The mentality for this type of computing need is just different; a special thanks to Brewster for showing me the way.
Back to the zombies… If you are a large incumbent or startup selling storage to support client server applications you are already dead, you probably just don’t know it yet. Modern companies don’t need what you are selling and the companies that have traditionally bought your gear don’t want new apps built on top of your old platforms. I know this is hard to believe, especially if you’re selling $25B in storage; but go ask Twitter, Facebook, Google or Amazon what percentage of their storage is done on these thirty year old architectures.
Mr. Zombie, don’t get me wrong, you’ll have a good business. You can still sell your legacy gear for legacy apps and still make billions; but your margin will compress, your growth will slow, and you’ll see new companies take the bulk of the budget for modern data centers. You will try to innovate by “buying your way an R&D budget.” Sand Hill will thank you, but you’re simply faced with the breaking of a compute cycle. Client Server did this to Mainframe; and there is still money being made in Mainframe, so as a zombie you will decompose slowly.
So what are you to do if you’re looking to build applications in this new world? Well, first and foremost you need to #DemandIncompatibility, which is Formation’s unofficial motto. The only way to leverage things new is to break the methodology of the past and start thinking like a thought leader.
If you start from scratch, ask yourself what you will need from your infrastructure? You will need: speed, efficiency, ease of management, multi-tenancy, massively scale out (not 32 nodes), hardware (HW) agnostic, etc. Oh yea, and by the way, everything is as-a-service and is based on a common API’s that S3, Swift, or Cinder supports. Things like media (flash/disk) are just the components. These are just a few examples of what applications to ask for from your infrastructure.
In this new world the grass is greener and the grey skies of the Zombie Apocalypse can be left to the data centers of yesterday. I’m so excited to be solving this problem, while not quite as cool as Michonne, at least we don’t need a machete to keep the zombies out of the new world of IT.