AWS re:Invent 2021- Day 1

Chloe McAree
4 min readDec 1, 2021

Wow, what a day! 3 hotels, 25,000+ steps and plenty of great technical discussions, today marks the first day of AWS re:Invent 2021.

This year I was lucky to receive the AWS All Builders Welcome Grant to attend Re:Invent in person in Las Vegas, Nevada. I’ve never been to a conference this size before and I’m super excited to share my experience of the event with everyone throughout the week through blogs like this one.

Before I get into the whirlwind that has been day one of the conference, I’ll start by introducing who I am.

I am Chloe McAteer, a software engineer from Belfast. I am an AWS Community Builder and a former Belfast City lead for Women Who Code.

Registration

Day one started with registration at the Venetian hotel, which involved picking up my conference pass and some AWSome swag!

Calm before the storm at registration

Break out sessions

With 1000+ sessions taking place throughout the week at re:Invent it can definitely be hard to select which talks to attend. However, one of the great things about this year is that all breakout sessions are recorded – so session conflicts are not a huge deal!

A few weeks ago I pre-selected and reserved seats for a few talks in particular that caught my eye. Here are the main highlights from them:

The first breakout session that I attended was on reliable scalability on Amazon.com, presented by Seth Eliot and Kieran Kavanagh.

This was a super insightful talk, it was great to hear the story about how Amazon.com went from a giant monolith that took 12+ hours to build, to thousands of micro services which are deployed independently, and in total are deployed more than 150 million times per year!

It was great see how their current micro/nano service architecture ensures that they can be independently deployed and scaled. This also allows feature teams to independently develop and improve certain aspects of the individual services/features, without having major downstream impacts (or waiting 12 hours for a build to run!)

One major thing that was at the core of the development of these micro services that all connect to one another, was how crucial it is to introduce a circuit breaker patterns for downstream dependencies. This ensures that you can handle failure fast and prevent throttling of the dependant services.

A great resource that was shared as part of the talk which can act as a guide for your architectural decisions to help ensure you are building reliable and scalable systems – check it out here.

Monolith, SOA & microservicees architecture

The second session I attended was by Donovan Fritz, a Network Engineer from Netflix, on how they are migrating to use IPv6 to enable hyper-scale networking. It was really interesting to understand the challenges Netflix started to see with their existing platform architecture that was using IPv4 and the impact it was having to their overall growth due to hard limits with ENI density.

It was also super awesome to hear how Netflix has transitioned into using IPv6 and some of the issues they initially faced, which lead to them creating their own custom transitioning mechanism to support backwards compatibility with older services that relied on IPv4, whilst still maintaining their overall security group preferences.

The final session I attended was on taking a prototype into production. This talk provided some super useful tips: firstly, to utilise serverless where possible. This allows you to get started very quickly at an extremely low cost whilst experimenting, but also setting yourself up to deal with sudden growth (and hopefully avoiding those extra costs!).

Another useful tip was to make sure you are thinking of security concerns right from the initial prototype – it can be very easy to “star all” IAM/resource permissions when trying to get something up and running, but ultimately this creates more work for you in the future when you take this into production. So it is always good to start with least privilege and increase permissions if need be in the future – but critically ensures that your project is always secure.

I plan on blogging throughout the conference — so be sure to check back here or follow me on Twitter to keep up to date with the ins and outs of re:invent and all the different events/sessions I attend!

--

--