AWS Summit 2019 - Tips, tricks, insights and socks
I always look forward to the Amazon Web Services (AWS) Summit and this year did not disappoint.
The AWS Summit 2019 was packed with interesting and insightful keynotes, breakout sessions and exhibitions demonstrating the latest and greatest of infrastructure and DevOps technology from AWS, their partners and clients.
The event was held in the ExCel London conference centre at Royal Victoria Dock in London Docklands. Once we had registered, we hit the Exhibition hall at the expense of catching the AWS summit keynote; this is a hot tip for anyone that goes in the future. I personally find the opening keynote is very sales focused; if you are willing to take the risk that you will miss an exciting announcement, the exhibition hall is relatively empty. This allows you to walk around at your leisure and talk with the different vendors and AWS staff, to find answers to the questions relevant to you and/or your business. Also, there is loads of free merchandise that runs out very quickly from t-shirts, socks (thanks Kong; I am wearing them while writing this article), red hats (from RedHat), coffees, free food, stickers, books, the list is endless.
We spent some time with partners we already use like Redis, Nginx and Pure Storage. As well as seeing an interesting AR experience from Deloitte; the duck factory! This AR application was adapted from a real-world client application used to train and demo production line machinery for the conference.
However, the most useful and insightful part of the exhibition was having the opportunity to talk with the architects from AWS. We got some useful tips and guides on the AWS AI/ML Service Sagemaker. I also got pointed at a wealth of training information around ML on the AWS site; split into tracks depending on your job/role, I.e. there is a developer track for me. This can be found here https://aws.amazon.com/training/learning-paths/machine-learning/ and in addition to that a free publication called Dive into Deep Learning which can be found here https://en.d2l.ai/d2l-en.pdf and covers the fundamental topics of ML.
We also spoke to architects on IoT and integrating AWS IoT Greengrass with Rasbery PI, Containerisation using containers using ECS, EKS and FarGate. We got costing tips for optimization our existing AWS infrastructure; which will directly save our customer money. There was loads of information on AWS training and certifications and partnerships.
We also got some exciting news from the Alexa skills guys; that in America they have rolled out in-application purchases for services, subscriptions and items which links with Amazon Marketplace. This is coming to the Europe over the next 6 months; we can already see how this can benefit our clients in the near future.
Once we had finished with the exhibition, we planned the breakout sessions we wanted to catch. They had a variety of sessions on various topics all running in parallel. The AWS event app on Android and iOS is useful for planning your day; although due to the number of people at the conference the mobile data network does grind to a halt. I managed to catch the following sessions:
Becoming a machine learning developer
This was heavily focused on AWS AI/ML Service Sagemaker; as well as going through the top-level progress flow for tackle an ML project. This is followed up by a case study from Lebara and how they had used ML to help combat call fraud. Sagemaker itself seems to have 17 common in-built machine learning algorithms, as well as being able to integrate with built-in frameworks such as TensorFlow, Chainer, mxnet, PyTorch etc as well as code written in C++ or R.
Buidling modern APIs with GraphQL
We already use GraphQL in many of our architectural approaches to aggregate data from multiple sources and support a microservice approach. This breakout session went through the basic of querying, resolvers etc as well as the history of GraphQL and a direct comparison to using a RESTful API. They also went into depth around the benefits of using GraphQL as well as how authentication should be handled with the request and authorization should be handle with the domain/business logic.
Modern application architectures
This was a very useful top-level view of the full modern architecture landscape which provides Secure, Resilient, Elastic, Module, Automated and Interoperability applications. This session covered Scalable infrastructure, microservice architecture, using AWS Lamdba, CI/CD pipelines, Blue/Green deploys, Containerization using ECS, EKS and FarGate
.NET in AWS
This was a useful session although; we were already up to speed with this. The biggest take away from this was use the AWS Toolkit for Visual Studio. Having used it myself I fully agree that it helps integrate .NET application with AWS, as well as allowing you to manage lots of infrastructure and services from Visual Studio as well as assist with deployment
The AWS Summit 2019; this year was dominated by AI/ML, which was mirrored in SXWT. This illustrates the drive from the big tech giants to build adoption in this area. Which is something I am very excited about and will ultimately end up changing all our life. If you get the chance to go to next year's Summit, I highly recommend it, hopefully I will see you there.