The AWSpalooza took me to Vegas for four nights, with thirty thousand or so other cloud-heads. Herewith notes and sparkly Vegas pictures.
Growth · The numbers tell the story: from 12K to 19K to 32K, and I don’t see any reason it’ll slow down. While the organization and logistics were formidable, obviously the work of seasoned pros, we’re getting close to the limit of what those venues can bear. I’m pretty relaxed about life, but had a couple of little claustrophobia flashes, when the crowds overfilled those huge hallways.
There are upsides: I ate like a horse and drank like a fish, but I bet I lost weight from the miles and miles (not a figure of speech) of walking from hotel to venue to venue. 24 hours away from the show, my feet are starting to feel less like undercooked hamburger.
Customers · There are two big tribes: First, the cloud natives, tiny to huge, who’ve never really thought of any other way to do computing. Then the much larger tribe just getting their toes in the water and figuring out what they’re going to have to change to get the public cloud’s upsides in cost and security and availability and durability.
The second group is bigger. I was talking to a guy from a British bank that sent fifty people for a crash course in the future. But he told me there was still a strong No-Cloud-Here faction, some in corner offices. “I can out-wait them” he said, “but in the meantime we have to get ready.”
Oops · I like to start the customer meetings (each day had many) with a question: “What’s not working? Tell us about your pain points.” And they laugh but then the ice is broken and you get a good talk going right away about the things that matter.
So we sat down with this one big insurance company (you’d recognize the name) and I asked the question; they looked surprised and started talking about the problems with monolithic legacy Java and lingering RPG and DB2 in corners of the business. I’d mixed up my briefing docs and hadn’t realized they were just starting the cloud migration, didn’t really have much in production yet. So I was embarrassed and apologized, but they said “No, this good, let’s keep going.” And actually it was, we learned things that they were going to have to watch out for and also some low-hanging fruit they can win with in the short term.
Secrets · When I was at Google, we couldn’t keep any — by the time IO rolled around every year, the press and bloggers knew pretty well what we were going to release. I’m not sure it did any damage, but it was irritating as hell.
AWS is a tight ship, relatively; we managed to surprise the audience with a couple of things, this year and every year. I totally don’t know why; if you listen to the AWS announcements, it’s obvious that customers have been looking at the new products, so the number of people who know is not small.
The Launch · I helped launch the new AWS Step Functions product. My role was small — flipping a couple of GitHub repos public, pushing a Ruby gem, publishing a spec — but enough to get me into the Launch War Room in a hidden corner of the conference.
Getting all the service pieces live on the net in sync with their keynote debut is not unlike a ten-player eight-dimensional chess match; I’ve never seen anything like it. I guess I have to be careful of giving away secrets here; suffice it to say, it was pretty groovy.
Speaking · I gave a session to an audience of a thousand and change; my first public appearance as an Amazonian, on my second anniversary here. It wasn’t as much fun as I had berating audiences about privacy in the time between Google and Amazon, but I do like speechifying.
re:Invent is speaker-friendly. At Java One, your talk was ripped out of your hands and edited by “professionals” who didn’t understand the difference between 1010 and 1,010. At Google IO, you got to keep your own talk, but you had to rehearse with, and get it approved by, Developer Relations people (like me) who ruthlessly stamped out bullet lists.
For re:Invent, they had professional editors, who were smart and helpful about style and branding correctness, but otherwise got out of the way. If there’s a re:Invent in your future, I strongly recommend getting a talk accepted; the Speaker Work Room is a haven of quiet conversation, free food, and strong Internet.
Next year? · If I still have this job, it’ll be hard to not go. I think people who are building into the cloud — even if, like me, they don’t like Vegas and don’t like crowds — should too.