This morning will be the last keynote for this year’s re:Invent 2018. It has been another great year for AWS with all the new technologies they are going to market with. In this blog post, I will be talking about some of the highlights from Amazon’s CTO, Dr. Werner Vogels’ keynote.
Dr. Vogels gave a deep dive into some of the AWS systems and announced several really interesting technical features, mostly focused on serverless technologies. Remember to register to our re:Invent Recap next Thursday, 6 December from 1.30 pm to get more details of these announcements!
Here are some updates:
The biggest announcements were around new runtimes for Lambda. While he started off with mentioning support for Ruby, the big kicker followed straight after when it turned out this was built using the new support for custom runtimes. It is now possible to build your own runtime for Lambda! Several partners have already come out with this so there is support for PHP, Erlang, and even COBOL.
The other big improvement here is support for layers, where you can now have shared layers. Which translates to being able to more easily share code between different Lambda functions.
Accessing your Lambda functions has also become easier, with Lambda being accessible from ALBs now, and the API Gateway supporting WebSockets later this year.
Lastly for Lambda, the Serverless Application Repository now supports nested applications. This means that you can reuse applications inside other applications.
AWS Lambda Supports Ruby
Use Any Programming Language and Share Common Components
Application Load Balancer can now Invoke Lambda Functions to Serve HTTP(S) Requests
AWS Serverless Application Model Supports Nested Applications Using the AWS Serverless Application Repository
Step Functions improvements
When step functions was introduced, it was as a way to tie various Lambda functions together. This worked well, but today we get a big improvement where Step Functions can connect to other AWS tools as well. This includes a varied range of services. AWS Batch for batch jobs, container services like ECS and Fargate for when you need compute that Lambda can’t handle, but also DynamoDB, SQS, SNS, Glue, and even Sagemaker for machine learning.
We have always been focused on delivering well-architected solutions, and have helped our customers by doing Well-Architected Reviews for them for the past year. A lot of this power will now be available for everyone, as AWS has released a tool that allows you to run this yourself.
This is a good way to see for yourself where your workloads can potentially be improved, and it also helps guide you into what it all means. Of course if you prefer someone to help you with this or want to follow up on the results you find, we are always happy to help out there.
AWS also released deeper integration with several IDEs and a managed streaming Kafka service.
written by Arjen Schwarz. Arjen Schwarz is an AWS Cloud Ambassador currently at the AWS re:Invent Conference in Las Vegas, Nevada.