I didn’t go to re:invent this year but it does not mean I cannot stay in sync with AWS latest and greatest – and today there were quite a lot of cool and amazing services and capabilities coming out either soon or already out there for developer to use and harness.
Obviously there are a lot of new services, capabilities and even on appliance in the machine learning field. And some amazing developments in IoT, security and new EC machines we can use with a click of a button. But as groundbreaking as those new shiny toys are I found myself excited about the ones that solve problems I have right-now, in my day to day, micro-service focused job.
I was happy to hear about how Amazon makes serverless computing even more powerful with Aurora serverless. By dividing data and database Amazon managed to create powerful, scalable and lambda useable database engine that is a real game changer for anyone working with functions as a service.
The other Lambda/scalable AWS specific database a.k.a DynamoDB was not left behind. Now (or rather soon) we’ll be able to replicate the data across regions with Global DynamoDB And now we have a data source which is not only scalable for one region but for as many as you need. And on top finally we can easily backup and restore DynamoDB in a managed way – just like RDS (in fact better than in some ways).
Another cool feature I’m waiting to get my hands on is Amazon S3/Glacier Select. It’s a solution to a problem I’ve always had but never knew could be solved this way. Storing files on S3 is only part of most solutions, in order to query the data inside S3 and effectively turn it to a simple but useful database.
But the two announcements that got me excited are the arrival of EKS – Amazon managed Kubernetes, we’ve been using Kubernetes in more and more projects in the last couple of years and finally we have a managed service that would help with the heavy lifting.
And finally a new service called AWS Fargate – which runs docker containers without an actual cluster, just like a new EC2 instance. I don’t know about you but I want to try this out, the possibilities of using a container and a single entity, not as big as a virtual machine (EC2) but has it’s own environment (unlike Lambda) should create another set of tools and patterns.