Musings on My First AWS re:Invent Experience: AWS re:Invent 2018
I wanted to get my thoughts together after all the hype and excitement had mellowed out a bit surrounding re:Invent 2018. Here are my thoughts to how you can get the most out of AWS re:Invent a week after it has already concluded. Scattered in here are also my thoughts and perspectives about the content. Only take them as such, I’m not necessarily an authority on AWS, I’m just a dude that loves high performing app dev/ops teams and their tricks of the trade.
Watch the Keynotes
If you do one thing, watch the key notes given, in full, sometime over the next couple months. For me, I loved seeing the direction AWS is going as presented, sometimes between the lines.
- Andy Jassy — http://2ndwatch.com/blog/aws-reinvent-2018-keynote-recap-wednesday/
- Werner Vogels — https://www.onica.com/blog/aws-reinvent-2018-recap-werner-vogels-keynote-event/
Continuous Education
Next, find 10 recorded sessions to watch on the AWS YouTube channel on topics that interest you or will help you address challenges you’re currently facing in your teams. There were so many good presentations given, you owe it to yourself to get through at least 10:
Serverless & Containers
Serverless technologies definitely appear to be staying. More than any other topic I saw, serverless technologies dominated the presentations and breakout sessions. The industry still is very fluid and rapidly evolving. Make yourself aware of all the serverless announcements and products that were given during the conference. My favorite serverless announcements were specifically around Lambda, API Gateway, and other similar technologies.
My favorite Lambda-ish announcements — https://serverless.com/blog/reinvent-2018-serverless-announcements/
- WebSocket support for Lambda functions
- AWS Lambda Layers
- AWS IDE integrations
- Improved Step Function integrations
- ALB support for AWS Lambda
- AWS Lambda Ruby support
- AWS Lambda Python 3.7 support
- AWS Lambda + AWS Kinesis Data Streams upgrades
My favorite Container-ish announcements — https://www.onica.com/blog/reinvent-2018-recap-containers-announcements/
- Open-sourced Firecracker virtualization technology (AWS Fargate)
- AWS CodeBuild and AWS ECR sources
- AWS CodeDeploy and AWS ECR sources
- AWS Step Functions + Workflow Automation for Containers
- Amazon ECS and AWS Fargate now support blue/green deployments via AWS CodeDeploy
- AWS Marketplace for Containers
I found that there are many companies worrying about the same thing we are at my current employer, that of vendor lock-in when it comes to working with serverless & container technologies. I found there to be a lot of different approaches, and even more solutions. The main thing that everybody seems to be worried about is that they will be stuck with a vendor if they decide to invest heavily on said platform. However, all of the vendors and solutions provided ways to keep business logic separate from routing and middleware logic. Also, there are many serverless abstraction frameworks (serverless.com, SAM, etc) being developed and that even currently exist that help alleviate these concerns, and even make local development easy and efficient.
Vendor lock-in does not scare me nearly as much as it did before going into this conference. The key takeaway personally is that as long as we’re correctly developing/architecting our apps, keeping business logic separated from the middleware/routing/auth logic, then the only thing we have to worry about is the middleware/routing logic if we move our code between cloud vendors.
With respect to containers, AWS is putting significant effort in to make them easier to work with in current AWS Pipelines (ECR is now a valid CodeBuild or CodeDeploy selection for source or destination).
It also seemed that there was less emphasis on AWS ECS and much more emphasis on AWS EKS and Fargate. It almost felt like AWS realized that the industry isn’t going with ECS as a container solution for enterprises. I can’t really back this up with any data, only observations and the lack of presentations on their original container solution.
DevOps & CI/CD
CI/CD is almost considered ubiquitous at this point in the industry. There were so many new variations being presented that it was very apparent that most high-performing IT teams were already doing the basics. I wouldn’t be surprised if we start hearing about tool-specific variations like GitOps, SlackOps, LessOps, etc.
Git and GitFlow usage were assumed across-the-board in all of the more prominent, packed sessions on development that I attended. Everyone just assumed that everyone was using it through their version control platform of choice. If you’re in the modern information technology industry, make sure you address this first for your teams as it sets up Continuous Integration from the outset.
More from monitoring tools and insights should and CAN be expected for our current services. In the expo, 1 in 3 booths seemed centered on new products providing data insights, monitoring, and observability in to services backed by DevOps workflows. Data seemed the name of the game for the future of data services. It was clear that products currently dominating the market have only scratched the surface of providing the required insights behind enterprise service meshes. We need to start expecting more from our current monitoring tools, and ditch them if they’re not giving us the visibility and functionality we need. Simply put, we can expect much more than we currently have.
Turnkey Solutions to Kill For
Their control tower, data lake with a click, and security dashboard, were awesome. I don’t know if they’ll meet the demands of a real enterprise, but for boilerplate functionality out of the box for startups, I would’ve killed for these 5 years ago. The details were:
- AWS Control Tower — automates the setup of a baseline environment, or landing zone, that is a secure, well-architected multi-account AWS environment
- AWS Security Hub — gives you a comprehensive view of your high-priority security alerts and compliance status across AWS accounts
- AWS Lake Formation — set up a secure data lake in days, not weeks
Machine Learning
During many of the keynotes and many of the sessions machine learning was emphasized in more ways than I can convey with words. The excitement, focus, and direction that AWS gave during this conference to machine learning was astounding. It made me realize that I am way behind the curve on knowing how to even use it.
Lack of Diversity
So regarding women in technology and diversity, nothing quite highlights the lack of diversity in our industry like a conference similar to re:Invent. Maybe it was just the venues I was in, but brogrammers were everywhere and honestly seemed to me that 5 of 10 attendees fit the stereotype. I know that’s a blanket statement that isn’t quite fair, but it was my observation.
I found that some of my favorite sessions were by women in technology. Here are two of my favorites (https://www.youtube.com/watch?v=HCCkVz25UU4&t=295s and https://www.youtube.com/watch?v=01ewawuL-IY&t=2108s). Some of the most bright, delightful, tenacious and talented colleagues I have at Magellan are women. I hope we can see more women get in to this industry and help crack it open to more diversity.
In Summary
But to wrap it up, my #1 observation is that this event is far more than a simple conference. I was naive coming to this event, even knowing it was in Las Vegas. When you cut through the marketing, the booths, the swag, the DJs, the parties, the casinos, the people standing on the strip corners flapping their papers at you, there really was a lot of content behind this behemoth. I can’t tell you enough about how much education was here. Continuous learning, improvement, and education was the #1 tidbit I pulled out of this event. Having Justin Kuss, Brian Lichtle, Trent Turner, and Vijay Gopal (all from Magellan Health) there as mentors for AWS didn’t hurt either.
