Following up on the recent Amazon Web Services conference, re:Invent, held at the Palazzo and Venetian in Las Vegas, the three monks who attended got together to chat about what we saw and the conference and what it means for AWS, developers and the rest of the industry. The conversation is lightly edited for content and for incompatible emoji.
sogrady: (Stephen O’Grady) Ok, we all attended re:Invent, so let’s start with the first question: what were your general impressions of the show?
monkchips: (James Governor) So glad I didn’t have to be in line for things.
The show is crazy busy now.
Every session was standing room only.
And don’t get me started on the pub crawl.
fintanr: (Fintan Ryan) The overall scale of the event. The content is good, but the sheer size of the event means its bursting at the seams.
sogrady: Agreed. The lines were so long in some cases I couldn’t even tell what people were in line for.
fintanr: On the product side though – my overall take is one of relentless incremental improvements since last year and Lambda all the things.
sogrady: Yes, the spread of serverless was notable, from the edge cases (literally) to Aurora and on and on.
How about you, @monkchips? What were your thoughts content-wise?
monkchips: My overall impression? A formidable company becoming even more formidable. I felt AWS made potentially industry-changing progressions in at least four markets, maybe five. It was like – here is the next billion dollars we’re taking, thank you very much. AWS is now thinking in terms of full stack platforms in the a16z sense.
sogrady: Interesting. What are the four or five markets you think are most impacted?
monkchips: We’ve all been working on IoT becoming a thing, and it felt like AWS is probably finally ready to start laying out some rails for people, with primitives that make sense. Thinking through what it means to manage swarms of devices for example.
The data group building services for managed Elastic, which also announced the Neptune database, also had the feeling of a market being made.
Neo4j has been working on creating a substantive market for graph databases, which map so well to so many modern software problems, for years. AWS builds Neptune, and had literally hundreds of customers in its preview.
sogrady: The AWS base is an inherent advantage, no doubt.
monkchips: Yep. Flywheels gonna flywheel.
So IoT, graph and search, and serverless as you’ve both mentioned. The serverless announcements were killer. The Lambda registry for discovering and sharing (and later billing) of serverless functions is potentially a huge market.
sogrady: The prevalence of serverless makes you wonder whether this is AWS’ solution for what used to be called the PaaS market.
monkchips: Yes – PaaS, without the opinion.
fintanr: Part of it perhaps. As @monkchips said without the opinion.
The serverless registry definitely grabbed my attention.
sogrady: @fintanr same. Enabling apt-get for serverless is likely to be more important than any single feature they’ve added.
fintanr: The spectrum discussion though implies they are happy for you to mix and match as much as you need.
sogrady: Ok, next question. Of all of the announcements, which do you think will be most strategically important and relevant to AWS and the market?
fintanr: I don’t think there is any one in particular – but if we look at a couple of segments, there are some highlights.
Containers and cloud native – Kubernetes is now a commodity, and Fargate is the far reaching part there. Particuarly when the Kube offering runs on Fargate.
monkchips: I thought the first rule of k8s club is nobody is allowed to make money in k8s club?
fintanr: Databases – again the spectrum, Neptune as @monkchips noted, but it’s the four pillars of offerings from RDS to Aurora to Neptune.
Except AWS @monkchips (edited)
sogrady: Don’t sell software, sell software that you run for people. Seems to be able to make a buck or two.
fintanr: Or as we often say – packaging, packaging, packaging.
sogrady: But anyway, you’re both:
sogrady: The most important strategic news was AWS’ work around the edges. From the app registry to the AWS Open work, the company is closing one of its few remaining vulnerabilities.
monkchips: Edge you say? When I talk about market making, the AI/ML stuff was interesting, particularly when tied to Alexa, voice services, translation, text to speech etc. AWS is leveraging adjacencies remarkably well for a company that thrives on building businesses and services around autonomous teams working independently. In the same way that Lambda will be used by everyone, so it feels like voice and the Echo business are transformative. Then throw in the Deep Lens camera designed for developers to play with (one of my favourite announcements of the show) with onboard AI. Image Recognition is the new hello world, so let’s get developers playing with exactly that.
fintanr: The AWS Open work is promising, but there is a lot to do there.
monkchips: Yeah let’s talk about that.
sogrady: No doubt. Way too much connecting of the dots is still required, and too many things haven’t been systematized.
But I look at it this way: previously, it was easy for AWS to essentially ignore open source because it wasn’t explicitly anyone’s job. Today, they’ve got a headcount of what, six or eight people now and are ramping up, all of whom are tasked explicitly with open source as a job.
So at least from an intent perspective if not execution, the signal is pretty clear: they want to get open source right. As they would define right, anyway.
fintanr: That is definitely true, but it’s a big metrics orientated org. The devil will be in the detail of the execution as you say. I’m mildly optimistic, but the trust will take time to build.
sogrady: And they may never get there. Or get to where people want them to.
monkchips: But of course it only really matters if the open source group can successfully run cover for product groups that want to make open source contributions. That decision will be up to product owners, not the OSS group. I feel that @iamstan made a really good point regarding open source. He went to a lot of enterprise customer sessions, and the AWS enterprise customer base is seemingly increasingly making open source contributions in their own right. Who is going to help foster those contributions?
sogrady: That’s an excellent question. If there’s one thing that would/could accelerate OSS interaction within AWS, it’s customer demand.
Even AWS critics acknowledge that the company is insanely customer focused.
fintanr: I do think AWS will get to where their customers want them to be, but perhaps not as far as some in the core communities want.
sogrady: But anyway, I guess my point is that more than any single product – be it serverless, database, global-spanning time services – a better relationship with OSS, its communities and the packaging around all of the above could be the most important single thing AWS could do.
Both for its own sake, and for eliminating it as a potential angle for competitors.
monkchips: I have been thinking about this a fair bit – the “AWS needs to contribute more” argument. At first glance it holds a lot of water. But another take is: where would commercial open source be without the cloud? AWS may not be contributing but Netflix is. If open source contributions are an externality of the reality of cloud, when folks like Comcast or Orange or a bunch of banks are contributing, that’s at least partly indirectly driven by AWS economics.
Bear in mind this thought is a trial bubble at this point.
fintanr: My take is that working with projects and making it easier for them to deploy and run on AWS is a big thing. Extra contributions are gravy.
sogrady: Fascinating point…that we don’t have time to explore in this particular chat.
Moving on…
We’ve talked a lot about issues of fragmentation, i.e. the difficulty that developers and end users have in making sense of the sea of options available to them generally.
We’ve also talked about this being a potential issue for AWS in the past given the breadth of its offerings. Did either of you get a sense that this is becoming a real issue for customers?
fintanr: Yes.
But more in a “I didn’t realise they did that too” fashion.
sogrady: Yes, I had several of those conversations.
“Wait, AWS has that?”
fintanr: There is actually a lot of the core common infrastucture people don’t seem to fully understand
Be that good use of IAM, CloudWatch, VPCs and so on.
monkchips: This is the post I am currently working on. TLDR? Issue yes, showstopper no. There will be kvetching, but delivery of high value services is worth more than stability. Pouring concrete on stuff makes no sense.
sogrady: Any ideas of how AWS addresses that? Increased levels of abstraction? More professional services style interactions?
monkchips: All of the above. The consulting industry should be very happy about it.
sogrady: The analyst industry too, it must be said.
fintanr: Abstractions and further education.
sogrady: Ok, next question: what did this show mean for AWS competitors?
Obviously re:Invent with its sheer volume of announcements and feature drops is never good news for competitors, but was this any more or less daunting of a show for competitors than in years past?
fintanr: I have been wrestling with this one – for most of their competitors more daunting.
Primarily due to the relentless customer focus – they now have a customer base that are willing to experiment, a lot, with new services.
sogrady: Why so, @fintanr?
fintanr: An enterprise customer base, that is.
sogrady: Ah, so the accretive aspects to their business become more problematic for competitors over time. Interesting.
fintanr: Precisely.
sogrady: Et tu, @monkchips? Was this same, worse or better for competitors than past shows?
monkchips:
fintanr: lol
monkchips: Definitely not good. AWS is showing no signs of slowing down, is making new markets, and like I say, doing some very very interesting work to finesse adjacencies.
sogrady: That GIF…yikes.
Setting aside the strategic, vendor sports considerations, if you’re a developer, what are you most interested in playing with from the new announcements?
fintanr: So much to choose from:
Aurora Serverless
Neptune
DeepLens
FreeRTOS
It’s more choose your area of interest and go play.
sogrady: For my money, it might be SageMaker. A fully managed ML service that basically abstracts all of the effort for playing with “AI” seems like a neat toy to play with.
fintanr: Yes definitely, it’s a far more than just a souped up Jupyter notebook.
monkchips: I will go with one of the sleeper announcements. It didn’t get much attention at the show actually but AWS announced AppSync, for building data driven apps with built in synchronisation. A few years ago I started talking about The Synchronized Web, with mobile services and peer to peer interactions across devices. I think AppSync could be a big thing in building the 2 tier architectures we’re seeing so much of. Parse showed the value of back end as a service for mobile developers. I think AWS will carve out a lot of developer interest there
Amazon Web Services, Inc.
sogrady: Right, I’d actually forgotten about that one. Too many announcements to process.
monkchips: +100
fintanr: Scripting Sumerian could also be a fun dalliance away from work for some people
sogrady: What I really want to know is whether Sumerian was a Snow Crash reference.
But anyway, heading down the home stretch here. Question: have either of you seen a recent conference that generated a higher number of inbound requests for meetings than re:Invent?
This is the highest density show I’ve seen since JavaOne a decade ago.
monkchips: Nowhere near it. AWS is a zoo. Absurdly target rich. On that note I was really happy we went off campus to a pub a few miles away. It was the best time at the show in many respects. The folks from Atomist joined us, and they said that RedMonk office hours at Atomic Liquors was the most relaxed they felt all week. Erica Brescia from Bitnami joined us too. Good times.
sogrady: All right, last question. We’re all in agreement that re:Invent has outgrown the venue, I think, so if you were AWS and could control where they held the next event, where would you have it?
Would you do it in a venue with a larger capacity, or would you attempt to break up the event into smaller events?
fintanr: That is a really hard question – we touched on this last year and I don’t know if there is a good answer.
monkchips: The revamped Moscone Center is the obvious choice. Or else AWS will build a purpose built city in the desert somewhere, its own version of Vegas. That will accommodate the coming re:Invents, which are basically 10 days of 24×7 announcements.
sogrady: And on that note, we wrap.
Thanks for your time, gents.
Disclosure: Amazon is a RedMonk customer, and paid for T&E to re:Invent.
