Day 3 at re:Invent

Day 3 for me started with an early breakfast. Serving breakfast to 13K+ people is a logistical feat made even more impressive when you see the scale of the dining facility. The hall is a gigantic warehouse-sized space on the lower floor of the Venetian whose far sides are hard to see from each edge. I somehow managed to be the first person at breakfast, due in large part to my East Coast body clock and an usher making a mistake in admitting me early. After walking for 10 minutes down aisles of wait staff directing me to a table like ramp service agents directing a plane, I managed to find a seat. Something about the wait staff lining the aisles reminded me vaguely of security at the Bangalore airport.

At 9 AM everyone was jammed in the hall for Werner Vogels keynote. He came out in his usual attire - concert t-shirt, jacket, sneakers. Werner is as close as we tech geeks get to a rock star. He started by telling us about the awesomeness of AWS, which was clearly a message targeted at new customers. He mentioned the importance of the partner ecosystem for monitoring and management (“"AWS is not just AWS - it's AWS plus partners services") before introducing the CEO of Splunk. Mr. Sullivan talked about how Splunk is going all-in on the cloud, although didn’t provide many specifics. To sum it up in one sentence, his message was: we love Amazon and are relevant now in the cloud.

Next up was the CEO of Omniphone, talking about their cloud-based music platform, and an executive from the Weather Company. The comment “I’m a weather geek. I think weather is amazing” struck me. Is there really such a thing as a weather geek?

It took almost an hour into the keynote before we finally got down to business. First up was the much anticipated announcement of the EC2 Container Service, which provides users an optimized scheduler for deploying containers. The announcement was made in conjunction with a short talk from CEO of Pristine.

Next on the list was AWS  Lamba, a new event-driven compute service for dynamic applications. The announcement was made in conjunction with Netflix, but I honestly never quite figured out the connection between Lambda and Netflix (maybe they just needed a way to highlight Netflix this year). The concept of Lambda both interested and horrified me. The interest is driven by this being potentially a new way to compose cloud applications on AWS. As it was being announced, I saw some social media chatter wondering whether this was Amazon’s equivalent of Google App Engine. The horror is driven by the thought that the poor application of this service by a company could easily lead to thousands of infrastructure-less functions running in the cloud like little pathogens. But overall I thought Lambda was the best new announcement of the conference.

Next up was the announcement of the C4 instance type family, which is a high compute family that uses a proprietary processor Intel produced for Amazon. The CEO of Intel came on stage to talk about the new processor and the importance of their relationship with Amazon. An interesting back story on this is that Jeff Barr told me that when he was writing the first blog post on this instance type, he Googled the processor to get more information and was surprised to get 0 results. Only after talking to someone on the product team did he realize this was an Amazon-only processor. Try to do that in your private data center.

After the keynote, everyone headed off to sessions or the exhibit floor. In talking with several attendees post-keynote, there were mixed opinions on the day 2 keynote. Some seemed to think this keynote was great, with one even believing this will be remembered as a milestone in the march of the enterprise into the public cloud But more people seemed to feel a slight sense of let-down. I suspect the let-down was driven at least in part by expectations Amazon might announce another price drop (nothing riles up the masses like a good price drop).

The CloudHealth team that worked our booth was looking forward to the 6 PM close of the exposition floor. They talked to hundreds of attendees and gave 200+ private demos over the course of a couple days. When the announcement finally came that the expo hall was closed, we headed off for a celebratory team dinner. Over dinner we swapped stories and lessons learned from the event. The overall feedback was that re:Invent 2014 had been a great success for Amazon, my company, and our individual technical and sales attendees. After dinner we split up, with one crew heading to the airport for a red eye, and another heading to the re:Invent after party.

There are questions & concepts I'd like to explore in future blog posts, including:

  • Vendor lock-in - How will companies balance the desire to leverage the power of AWS's new services with their corresponding impact on vendor lock-in?
  • Future of IaaS - With an increasing array of PaaS services like Lambda, what is the future of IaaS for building cloud applications?
  • State of the cloud - With Amazon producing more revenue than the next 14 cloud providers and the private cloud movement faltering, what will the future be for the cloud?
  • Coopetition - As Amazon responds to its growing customer base, it's finding itself getting into businesses many never expected (e.g. CodeDeploy, CodeCommit, CodePipeline, Aurora). How do they balance the needs of their customers with their desire for maintaining a balanced and committed ecosystem?

That’s all from re:Invent for now. It was a great conference, made possible by great preparation and execution by Amazon and its partners. See you all next year.


Related posts: Day 1 at re:Invent, Day 2 at re:Invent