A couple of weeks ago I was in SF at Google Cloud Platform Next 2016, an event designed to show Google is serious about competing in the public cloud market against Amazon Web Services, Microsoft Azure and IBM SoftLayer. The market and even some Googlers have been questioning the firm’s commitment to cloud as a line of business, so the company felt it needed to make a strong statement. Google duly totally over-rotated on personnel-as-proxy-of-seriousness on day one by having not just new cloud lead Diane Green, but also Sundar Pichai, CEO of Google, and then for good measure Eric Schmidt, Chairman of Alphabet all keynote. The session duly considerably overran. Never good on conference chairs. Maybe better to have just Diane next year.
I actually missed the day one keynote, which may be one reason I came away with a different overall impression from many commentators, who were seemingly underwhelmed.
The Day Two keynote was serious, committed, intelligent and challenging. The structure followed a nice pattern from high level, to low level details.
Google made the clearest commitment so far by any public company, let alone major cloud vendor, to renewable energy, of which more later. But TLDR Thanks Google, I have 3 children, and if Google scale is carbon positive that’s good news for every living thing on the planet, including my little ones.
— Jake Kaldenbaugh (@Jakewk) March 24, 2016
Google’s Neils Provos then went deep on its layered approach to information security, an impressive framework should pressure competitors to up their game. Google security provides an excellent set of approaches for delivering on security in the Cloud era. The geopolitical/industrial scene isn’t getting any easier, and information governance in the cloud is going to be a defining issue of our time for governments and businesses. On the flight out to Next I read an article in the FT about military contractors moving into the enterprise security market. Frankly any enterprise that calls BAE or Raytheon before it looks to Google about how to better secure information assets in the Cloud era is doing it wrong.
fan boy moment. it's eric brewer!
— Tax Doge (@monkchips) March 24, 2016
The final talk by Eric Brewer – he of the ubiquitous CAP theorem that drives so much modern distributed systems design – made it abundantly clear that Kubernetes is not just window dressing to Google’s own internal Borg architecture, but rather represents a pretty fundamental change in Google’s cloud posture – the company is finally learning to play nicely with other children, although it still has plenty of work to do in that regard. Google now leads a project that is crushing it on Github, with outside contributors, and it feels good about that. See this post from Fintan for more. It wants to build a successful ecosystem. But doing so takes time, commitment, attentiveness, listening and a degree of humbleness (this last one doesn’t come easily to Google).
There is no company more Cloud Native than Google. The company literally invented this stuff. Defining Cloud Native Brewer pointed to development that expects cloud resources to just be available, “an infinite number of machines”; other characteristics include containers for packaging and isolation, with a micro-services orientation.
But the question remains – can Google package up its experiences and platforms in ways that makes them easily consumable by enterprises and startups? The best packager in any tech wave wins.
It was particularly noteworthy therefore to see Google’s positioning of Kubernetes with respect to Docker during Brewer’s talk. Containers is nothing new, Google was making open source contributions as far back as 2006, in the shape of cggrounds. But Google’s tooling was designed by the best engineers on the planet for the best engineers on the planet. It works but it’s very far from being easy to use.
“Docker came along and did a better job of the packaging, it does a nice job of how you handle libraries”
Which sort of sounds like faint praise but isn’t. Docker utterly killed it with making containers a first class citizen for software development, and Google knows it. More than many other competitive vectors Google is chastened by Docker. Google wants to make sure that it benefits from The Docker Pattern as much as, if not more so, than Docker does.
That said, Google made it very clear that while it sees Docker as moving the state of the art forward with containers for development, it believes it can do a better job managing containers in production with Kubernetes. Google plans to niche Docker. The contrast was ironically enough heightened during the keynote, when Docker dropped its new Mac and PC clients, making containers even easier to use on the desktop.
Bottom line is Google wants to win those Docker workloads, but needs an ecosystem to do so. A recent Docker survey shows Amazon EC2 Container Service is already a natural target for these workloads.
While Amazon has dominated the first couple of rounds of the Cloud wars, this is going to be a long game. Enterprise workloads have only migrated to the cloud at the edges. Core transaction systems remain on prem. On that note Ron Harnick of Scalr said Next wasn’t boring enough
Where’s the bank that runs mission critical operations on GCP? Where’s the retailer that can run transactions faster on GCE than on EC2?
It’s the infrastructure, stupid. I mentioned above that many commentators were unmoved by the event. Of those I thought this post – Google’s Scalability Day – by professional curmudgeon Charles Fitzgerald was great. As so often lessons from history are particularly valuable:
In May 1997, Microsoft held a big press event dubbed Scalability Day. Microsoft was a relatively new arrival to enterprise computing and was beset by criticism it wasn’t “enterprise ready”. The goal of the event was to once and for all refute those criticisms and get the industry to accept that Microsoft would be a major factor in the enterprise (because, of course, that was what the company wanted…).
Microsoft at the time was an extremely engineering-centric company, so it processed all the criticisms through a technical lens. Soft, cultural, customer, and go-to-market issues were discarded as they did not readily compute and the broader case against Microsoft’s enterprise maturity was distilled down to the concrete and measurable issue of scalability. The company assumed some benchmarks plus updated product roadmaps would clear up any remaining “misunderstandings” about Microsoft and the enterprise.
The event was a disaster and served to underscore that all the criticism was true. It was a technical response to non-technical issues and showed that the company didn’t even know what it didn’t know about serving enterprise customers. Internally, the event served as a painful wake-up call that helped the company realize that going after the enterprise was going to be a long slog and would require lots of hard and not very exciting work. It took over a decade of very concentrated focus and investment for Microsoft to really become a credible provider to the enterprise. Enterprise credibility is not a feature set that gets delivered in a single release, but is acquired over a long time through the experience and trust built up working with customers.
I couldn’t help but think about Scalability Day while watching Google’s #GCPNext event today. After telling us for months that this event would demonstrate a step function in their ability to compete for the enterprise, it was a technology fest oblivious to the elephant in the room: does Google have any interest in or actual focus on addressing all the boring and non-product issues required to level up and serve enterprise customers?
As is their norm, Google showed amazing technology and highlighted their unrivalled infrastructure. And they have as much as admitted they’ve been living in an Ivory Tower since Google Compute Platform was announced in 2012 and “need to talk to customers more often”. Recognizing you have a problem is always the first step, but beyond throwing the word “enterprise” and related platitudes around, they did little to convince us they are committed to travelling the long and painful road to really serving enterprise customers.
So much all of this. Google doesn’t need to change its engineering. It needs to change its posture. It needs to be open and be seen to be so. It needs to give itself permission to come across as more human. Enterprises want to work with people that are like them.
The Cloud posture isn’t collegiate enough yet, although it is of course somewhat academic. Google is about the New Applied Science.
I think one obvious solution to a perceived arrogance issue is to focus more on partners. It was noticeable that Google didn’t feature partners on the main stage at least not on day two. It could have gained some kudos for example by featuring Red Hat talking about Kubernetes. We have seen a marked change at Google over the last few quarters, partly driven by infusions of new blood, but also in Google’s experiences working with outside firms – notably Red Hat. This didn’t come across quite as strongly at Next as it should have done.
But the narrative is there, waiting to be packaged. Over the next couple of years we will see Next become more about the ecosystem and less about the platform.