GIFEE is an acronym for Google Infrastructure for Everyone Else, and it is a term that has become a reference for open source equivalents of Google’s technology since it was unveiled way back in late 2015. Now we all know Kubernetes is based on Borg and was built to help us benefit from their years of experience with containers. We also know Prometheus is another part of that puzzle that the enterprise has found very beneficial and while both tools are great, do we really need to be wearing wristbands that say “what would Google do?” The reality, however, is that Google is so much ahead of everyone in the space and so aware of it, it’s almost like they’re looking down on the rest of us for being so technically backward. And that may well be the problem with GIFEE.
Kubernetes and TensorFlow
Google has never been especially “enterprise-friendly” and has always been more driven toward leading the way with cutting-edge technology like AI, rather than focusing on what their enterprise customers actually need. This is evident by the fact that even though Kubernetes is at the heart of modern day digital transformation, Google Cloud Platform is still a distant third in the public cloud segment. What this boils down to is the fact that enterprise customers would much rather use something that works for them and they’re familiar with than something that’s cutting edge or being used by Google in production.
Another good example of Google’s approach gone wrong is TensorFlow, which is a Python-friendly open source library for numerical computation that makes machine learning faster and easier. Now while it is by far the most sophisticated open source machine learning platform on the market, where it lacks is a large ecosystem of partner building tools. These should ideally include a large number of independent software vendors, tools, widgets and applications that others can use to build on. So while Google has a better product, Microsoft and AWS have done a better job of building out an ecosystem of Big Data analytics and AI-based independent software vendors (ISVs) and channel partners.
Google under Greene
VMware co-founder Diane Greene was appointed Google Cloud CEO roughly three years ago to help usher in the next stage of cloud development and there’s no doubt she made big changes and lots of progress before leaving the company last month. During Greene’s tenure, Google increased its annual capital expenditures from $10 billion to more than $13 billion and went on a hiring spree —Google Cloud hired more people than any other group at Alphabet over the last two years. It also got some key customer wins over Amazon such as Salesforce and The New York Times and built out several important functions for selling to enterprises, including professional services, training, and marketing.
Unfortunately, however, while both Microsoft and AWS were actively wooing partners and acquiring startups, Google under Greene were missing out on major acquisitions like Red Hat, GitHub and MuleSoft. IBM purchased Red Hat, while Microsoft bought GitHub, and Salesforce acquired MuleSoft. Reports suggest that while Greene wanted to buy GitHub, Google CEO Sundar Pichai was less enthusiastic and unclear as to why Google would spend big money to get into the market for developer tools. The report also claims that Google’s bid for GitHub, whose cloud software lets programmers collaborate and share code, came in at just under $6 billion, and it declined to raise the price after being told of Microsoft’s offer.
New captain at the helm may not be yippee for GIFEE
According to the same report on CNBC’s website, there had been considerable tension and disagreement between Pichai and Greene. Two former employees described a recent disagreement between the two over a controversial contract with the U.S. Department of Defense. Some speculate that this is probably why Google Cloud now has a new CEO — Thomas Kurian, who has spent the past 22 years focusing on Oracle’s database business and mostly ignoring the cloud. In fact, while Google Cloud definitely has problems, they pale in comparison to Oracle’s cloud woes, which include something called an anti-Oracle ecosystem.
Oracle as an organization was largely dismissive of the cloud at first and chief Larry Ellison infamously called it a “fad.” To make things even worse, just like Google, Oracle is famous for ignoring the need to create a large portfolio of related services to make every aspect of building applications easier. While Google has finally figured out it needs to have a wider portfolio, Oracle is still emphasizing its core database instead of the surrounding app development ecosystem.
Changing Google’s culture
Kurian brought Oracle Cloud into the mainstream and even prevented Oracle itself from being disrupted out of existence, but Google is going to be a different kind of challenge. Allen Falcon, CEO of Cumulus Global, was quoted as saying, “Google’s culture is not necessarily top down. It is more collaborative. He may struggle in getting Google to change direction. He is going to have to engage all the different layers and stakeholders at Google to make things happen.” What this means is it isn’t easy to change Google from what it is today, which is an engineering machine loaded up with Ph.Ds. in artificial intelligence that is developing the technology that will define the future.
Unlike what Nadella did with Microsoft, Diane Greene wasn’t able to change Google’s core identity from what seems akin to the enterprise version of nerdy Sheldon from “The Big Bang Theory.” This is because while AWS and Microsoft are geared toward serving customers and responding quickly to their requests, Google touts its own technology (like GIFEE), selling what it thinks clients need. Thomas Kurian definitely has his work cut out for him, mainly in the form of changing that Google culture that says “here this is how we do it, just follow us and shut up.” He also needs to build those channels and partnerships for Google Cloud to be successful. What Google needs right now are partners to tie every application in the world to a Google frontend.
GIFEE: Just too much for most people
The GIFEE concept itself is questionable, to say the least. This is because using Google infrastructure for a lot of people would be like driving the Starship Enterprise to work. Though it would be a lot of fun, the point is to get to work and get your work done and not to learn how to operate a spaceship. This is precisely the point Google doesn’t seem to get and while there’s no doubt we’re better off building on containers, we definitely don’t need to do it like Google with all their fancy, super-configurable tools. Instead of GIFEE, what we do need, are tools that do their job, get out of the way and leave us with enough time and patience left over to build a business.
Featured image: Shutterstock