While I was an analyst at Gartner, I co-developed the concept of “web-scale IT” based upon what I had learned from researching as well as talking to individuals working at several of the industry-leading FANG (Facebook, Amazon, Netflix and Google) companies. At the time, I envisioned web-scale as the next step in DevOps evolution as DevOps hadn’t seemed to have gotten around (at the time) to talking much about required changes in application architectures (microservices, etc.) nor how data center facilities should be optimally constructed. Yet, when I would present on the concept at various conferences, I was (I think usually) careful to state that I wasn’t recommending that enterprises blindly copy the approaches to IT represented by this illustrious group of unicorns, but rather I saw the FANG group’s capabilities as the new benchmark as to what was possible provided you were willing to make dramatic changes not only with respect to technology, but also processes, etc.
Since that time, another acronym has entered into the industry lexicon – GIFEE, or Google Infrastructure for Everyone Else. Alex Polvi, CEO at CoreOS, in a blog describes it as “a style of managing infrastructure where you can pull the plug on any server at any time, and the apps keep running — the way hyperscale companies like Google operate, with infrastructures designed for maximum robustness, scalability, security and reliability.” I first came across it in one of Craig McLuckie’s presentations (see GIFEE Google infrastructure for everyone else). Craig, as many in the cloud native world know, is a former Google Group Product Manager and is one of the original founders of the Kubernetes project. In his presentation he argues a case that enterprises can be more like Google by leveraging technology that has its origins from FANG and FANG-like companies (such as Docker and Kubernetes) in addition to the adoption in general of open source software backed and supported by a broad-based community. Gareth Rushgrove, who has been active in the DevOps community, does a nice job in his presentation of “The Two Sides of Google Infrastructure for Everyone Else” of wryly arguing both for and against the GIFEE concept. In essence, though, the negative perspective takes the position of enterprises are not like Google and thus what is normal for them may not be suitable for your IT organization.
Ever since joining VMware, I have been returning to this GIFEE/web-scale IT thread in my mind. Is it reasonable to assume that enterprises can become like Google or any other member of FANG? I know while at Gartner we made some pretty heady projections in terms of web-scale IT adoption. In fairness, I included DevOps as an integral component of the concept enabling me to claim that elements of web-scale were in fact being employed, but full-on Google in the enterprise? Given the interactions I’ve had so far with many organizations within VMware’s client base, I have to admit not so much (at least yet). Some readers of this post may draw the conclusion that I’ve been primarily interacting with what we used to call Type “C” enterprises (or technology laggards) at Gartner, but that’s not the case at all. Many of these firms are a) very sophisticated and b) in competitive industries and thus are responding to radical digital transformation.
So, what is the point of this rapidly becoming a “tl;dr” blog post? I see VMware picking up the mantle of the champion of GIFEE but not necessarily in the manner intended by the original proponents. The reason that I state this is that GIFEE in it’s pure form I think “breaks” most everything in traditional IT shops. It breaks huge investments in infrastructure, tools, processes and of course demands a commensurate investment in culture change. We can debate that these are sunk costs and therefore should not influence future purchase decisions, but that’s not always the economic (if not political) reality that many in these more traditional IT environments find themselves in. I believe that the infrastructure investments that VMware is making in technologies such as VMware Cloud Foundation, Kubernetes (Pivotal Container Service), Docker (vSphere Integrated Containers) and Functions as a Service with Project Dispatch are delivering on the spirit of GIFEE if not representing the exact technology stack. We’re not forcing IT organizations to disrupt everything they’ve done before in order to take advantage of the new “cool kids” stuff. More importantly, we are focused on making “infrastructure-as-code” an increasingly simple checkbox option anywhere, whether that “anywhere” is a cloud, traditional data center, or even out at the edge. In addition, we are committed to providing a native experience for how developers interact with that infrastructure, inclusive of tools and APIs, whether that’s native Kubernetes, Docker, Cloud Foundry, or whatever will gain traction in the future.
And then there’s what I call COFEE (Cloud Operations For Everyone Else)! Being someone with a management background, I may in fact be more excited with some of the management (and security) tools that we’ve recently delivered as well as plan to offer in the not too distant future to support cloud operations requirements. Enhancements in vRealize Operations 6.7 and vRealize Automation 7.4 as well as others that have just shipped are designed to enable the monitoring and automation of increasingly complex environments without having to totally learn a new management way of work. We’re also delivering advanced analytics and more intuitive workflows (and UIs) so that not everyone on your IT operations staff has to be an SRE (site reliability engineer) ninja. Don’t get me wrong, I’m still a big DevOps/SRE fan, but not every enterprise can staff developer-capable talent within their IT operations organization. Want to support private and (and increasingly) public clouds with the same set of tools? Check. And we’re building out an impressive collection of products that will enable you to do this not only on premises, but also using a SaaS-based paradigm further reducing some of the potential skills impediments (note: the most recent release of vRealize Operations enables the more rapid onboarding of our Wavefront service enabling quicker time-to-value for the DevOps set).
So, to net this out … I see VMware’s combination of leveraging both open source as well as traditional technology as perhaps being a real-life embodiment of the Pareto Principle with respect to enabling traditional enterprise IT departments to come closer to delivering (some of) the rarified capabilities of the FANG companies – without some (or most) of the pain. And because of this, the odds of success for these organizations attempting to improve their IT delivery capabilities will I think be higher. So, check out VMware’s GIFEE-oriented deliverables and while you’re at it, maybe grab some COFEE from us as well!
