Home > AWS, Cloud, re:Invent > AWS re:Invent 2017: The Show Preview

AWS re:Invent 2017: The Show Preview


Scaling Up and Out

re:Invent has turned into a vast conference, dwarfing many other IT get togethers, 50,000 people expected with an unmanageable schedule spread across 5 hotels. Even if you can walk quickly, its at least 30-45 minutes walk between quite a few of the venues.

This is the first time I’m attending re:Invent so the rush of excitement of a new event is invigorating for me.

I’d love to meet up with anyone who is here so please get in touch via @julian wood

AWS seems an unstoppable machine and that’s not in any way to say it should be stopped. Its rare a single company rises so spectacularly to create a major new part of an already established industry. 10 years ago this new thing called EC2 was a curiosity, now its old hat as services like Lambda become the new compute engine.

Why I’m here

I’m coming to AWS self funded so don’t have a company/vendor agenda I need to follow, I can truly see what interests me! My plan is to first of all see what’s new. This isn’t just about announcements although those are important but more how AWS develops into a service provider specifically for enterprises. Cloud is a new way of doings things yet enterprises have been doing things in a particular way for a long time, sometimes for good reasons and sometimes it is plain old toxic heritage IT!

I’m also investigating how enterprises change themselves to do cloud properly, taking advantage of the flexibility and scalability of public cloud and also how AWS updates itself to service enterprises better. This is a complicated dance which needs a lot of back and forth fancy foot work from both sides. I’m expecting more enterprise features, plenty of compliance, governance and security as a service. The recently announcement of PrivateLink as a new way to connect your VPCs to AWS services with more control is a nod to enterprises that’s don’t want all their stuff near the internet by default. AWS is working hard to ensure enterprises can take advantage of all the clever cloud stuff they have to offer.

Secondly I’m watching the serverless space very keenly and expecting a huge focus on Lambda. I’m not being hyperbolic by stating the serverless/FaaS pattern is the future of compute. A new cloud operating model where all infrastructure is abstracted, business processes as pure code without restrictions on scale, billed per invocation giving you financial super-powers.

What could be coming?

AWS releases new features and major updates to existing ones in an unprecedented manor. Just last week, more than 30 announcements were made and that’s before re:Invent and the big picture announcements likely reserved for the show. Make sure you subscribe to the excellent Last Week in AWS mailing list by Corey Quinn

At this stage its a guessing game as to what they might be, but certainly more IoT, ML, AI and buckets of enticing delicious serverless. I’m expecting some Blockchain as a Service and likely Kubernetes as well.

Amazon is also a remarkable company in the way it is able to erm. reinvent itself. This isn’t just an AWS thing but AWS certainly helps all of the rest of Amazon be incredible agile, be able to continually experiment, use its own platform as a sensing engine, rely on real data to make decisions and cannibalise itself. Lambda has been built to cannibalise EC2.

Time to delve a little deeper…ready?


Entering the Enterprise

Despite much adulation, AWS does however have headwinds. Paying enterprises have been very slow to adapt to clouds. Sure there are big names doing some clever things but most enterprises are still dabbling around the edges and working out how to actually do public cloud. What to put in what cloud is the question enterprise architects are continually asking. Public cloud isn’t actually as flexible as they might have hoped. Spinning up workloads that span multiple public clouds like Amazon and Azure is a rare occurrence. Picking a cloud for a particular application or workload is making more sense. This makes the decision often a little more serious as it isn’t as easy to change your mind later.

Enterprises are still very much attached to their own data centers, part of it is security and control which is fading but the main barriers are existing sunk costs in data centers that they may as well sweat and also the dramatic process changes they need to move to public clouds. Each public cloud has its own operations. Sure, its easy to spin up the same Nginx web server or MongoDB database on any cloud but how do you change your tooling to rather use Lambda and DynamoDB?

On-Premises and IoT.

Enterprises that are fully embracing and moving to public cloud may still have on-prem things they need to work with but AWS can’t manage things internally. The nature of the internal things will certainly change, we’re not only talking about connecting to an on-prem VM, SQL DB or mainframe. On-prem things will be sensors, scanners, printers, all manor of things that will physically connect the people working, the sensors gathering, or the systems managing buildings, factories, train stations or oil rigs. How will AWS more effectively reach into these environments and link them up with the cloud. I’m expecting the idea of Greengrass will be used to land and expand this vision, running more Lambda on-prem. Lamdba@Edge was announced last week which despite the Edge part of the name really means better routing options to fire off Lambda functions depending on which edge location you use to hit the AWS cloud. So, Australian users get data back more quickly from Lambda functions running in Sydney and German users hitting the same overall service get data back from Frankfurt based Lambda functions. The “Edge” is still in the AWS cloud, not pushed further towards the rest of the world. Could this change? Where does AWS see the future Edge sitting particularly regarding IoT?


I’ve already mentioned quite a bit about serverless and Lambda and I expect there will continue to be a huge focus at re:Invent. AWS recently announced Lambda can run functions with 1.5Gb memory and run for 30 minutes rather than only 5 before. That’s quite an update, these microservices are allowing more macro use cases. There will likely be updates to the Lambda front door, API Gateway, and more back end AWS services linked in. I’m sure I can’t even imagine the use cases that are possible, I hope we see some awesome!

Open Source

AWS is making a lot of noise about open source recently. AWS has traditionally been a major user of open source however the feeling is it hasn’t really contributed as much back. I can see AWS making this a much bigger focus. Part of it is certainly being seen as a better software citizen and sharing its toys but its platform is becoming a standard in some areas and it must realise it should foster further innovation to increase the platform coverage.

The Fun

There are surely going to be some headline grabbing fun announcements. AWSs scope is so broad and it covers so many areas. Can we see some fun, interesting and out of this world demos in AI, ML, game production, IoT, finance….in fact anything? Sure, the infrastructure coolness is there, heck, you can plug a truck into your datacenter to help migrate your data to AWS but what are they cooking up that’s way beyond coolness?

The Take Over?

Looking at the bigger picture, AWS is just a vehicle for Amazon at large. What other industries can Amazon cannibalise? With such a powerful and extensive computing platform Amazon can really go for anything. What’s next? Which company executives are going to start sweating during the conference keynotes as they realise Amazon is now gunning for them? The financial sector is ripe for some ripping up. Entertainment with VR? The Medical Money Machine and Big Pharma? Any industry is up for grabs!

Can’t wait to get stuck in and find out more!

  1. No comments yet.
  1. No trackbacks yet.

Time limit is exhausted. Please reload the CAPTCHA.