AWS Summit New York: From Infrastructure to Platforms
Last week we made our final stop on the “road to re:Invent” for the 2015 AWS Summit in New York City. As usual, the Big Apple was a hot, sticky, mess of goodness complete with exciting announcements from Amazon CTO Werner Vogels, who introduced AWS Device Farm and the Amazon API Gateway to a packed crowd.
I won’t bore you with all the nitty gritty details of his keynote address (you can see it for yourself here). Instead, I want to talk about something important that Werner said which I don’t see getting coverage anywhere else.
From infrastructure to platforms
“We’ve moved away from Infrastructure” It was right there 30 minutes into his keynote address. If you weren’t listening closely you might have missed it.
Werner, who was wearing a “Rage Against the Machine” t-shirt under his patented black blazer, was very quick to qualify the statement. Literally, as soon as it came out of this mouth, he immediately followed up by saying that AWS still indeed provides infrastructure but that they now offer platforms. Platforms that let a business focus on their products–or rather on building products on top of AWS.
So was this a slip of the tongue, or was AWS’s own CTO telegraphing a new way that they themselves are viewing their own cloud?
Personally, I think it was the latter. Let me explain why.
Avoiding the race to the bottom
After attending a dozen or more keynote addresses given by AWS executives over the past three years, one thing is very clear: AWS is desperate to distance itself from the meme that it’s in a race to the bottom with Microsoft and Google over who can provide the lowest cost for infrastructure as a service. The counter-theme for the last year has been that AWS provides more depth in any major infrastructure category (compute, storage, and databases) than any (and probably all) of its competitors–that’s it’s essentially in a class of its own.
It’s hard to refute this when you look at the sheer number of new products that AWS continues to introduce. The AWS cloud is not just wide, it’s also very very deep. However, despite this, the race to the bottom meme continues to persist and flares up every time one of the big three changes cloud prices. This hasn’t been a big issue in the past but now that AWS financial performance is being reported separately by Amazon, Wall Street is now listening to the noise in an attempt to find the signal.
Thus, the turn towards “platforms” in Werner’s remarks.
Take a look at the product/service announcements AWS has been making over the last year through the platform lens. If you do, a new picture comes into focus.
The picture shows AWS developing two end-to-end technology platforms: one aimed at web based applications and the other clearly focused on mobile applications which continue to see skyrocketing levels of usage due to the adoption of smart phones and tablets. These are not simple PaaS offerings, but rather full lifecycle platforms with tools aimed at every step, from application development, to deployment, to ongoing management.
Plugging holes in the platform
Nine years after launching, AWS now has the breadth of infrastructure offerings to support almost any workload (web, big data, HPC, etc) that someone could throw at its cloud. Lack of the right type of infrastructure is rarely the friction point today for companies to adopt the AWS. So naturally we see AWS focusing on the other areas that prevent migration to their cloud: namely development, deployment, and management.
By plugging these holes, AWS continues to look a lot more like end-to-end web and mobile application platforms. And that is exactly what they are doing.
Last November at re:Invent we saw the introduction of AWS CodeDeploy, CodeCommit, and CodePipeline. These are services were aimed at the “last mile” of application development—an area dominated by a fractured set of third party developer-oriented tools for source code control, continual integration and deployment.
In NYC a few days ago, it plugged another hole in the application development value chain with Amazon API Gateway, which does for APIs what Elastic Load Balancer (ELB) does for load balancing traffic.
On the mobile platform side, it was almost exactly one year ago that AWS introduced a huge set of mobile services that took the sting out of developing mobile apps. AWS Cognito, Mobile Analytics, and the Mobile SDK all made simple work out of the complexity surrounding the management of mobile identities across devices, measuring usage, and talking to AWS services from within iOS and Android apps.
Last week in NYC, AWS plugged a major hole in its mobile platform with AWS Device Farm—a comprehensive mobile device test harness that allows developers to run their app across a dizzying array of physical mobile devices.
Time will tell
Only time will tell whether or not this platform view of the world will be adopted by the companies and developers that use AWS. However, more and more I’m hearing CTOs and CIOs make the the decision to “go with AWS” as a whole rather than deciding to adopt just one or two core infrastructure services like they did a few years ago.
One thing remains certain though, Werner rarely ever says something he doesn’t mean.
See you at re:Invent.