It’s been one year since we last reported from AWS re:Invent Las Vegas. Last year we saw a couple of exciting announcements in the infrastructure space and new feature launches, most of them around SageMaker becoming the de facto stack for developing and launching ML-based solutions. For this year, since we’re in the middle of a pandemic, AWS has decided to deliver the re:Invent experience fully remote for everyone at home, featuring the usual content such as sessions, hands-on labs, demos, and the usual long-format keynotes. We still miss getting swag for the team at the expo, though.
AWS Lambda, one of the key services for Serverless workloads, is getting packed with new features and announcements that will make you reconsider it if you are not using it to its full potential. Right now, AWS Lambda offers better management capabilities, desirable price reductions, support for containers, and lastly, more compute capacity for running compute-intensive tasks for your serverless workloads.
Let’s review them one by one briefly.
Still in preview, yet exciting is AWS Proton, a service built for facilitating the management of infrastructure and code deployments. It was created specifically for Serverless workloads and any other container use case.
AWS Proton helps your core infrastructure and cloud teams define templates that can be later be utilized by your development team by following the AWS best practices, including those of the AWS Serverless Lens as part of the AWS Well-Architected recommendations.
Proton can integrate into your existing CI/CD tooling and other observability tools through a streamlined process.
AWS Proton will do all the work for your teams. This includes provisioning, setting up any software or third party observability tools, and AWS native solutions. With this, your development teams can focus on the application and container code and continue to build features to bring more value to your customers. All deployment related tasks such as version management, monitoring, rollbacks, and production deployment are relegated safely to AWS Proton under a single pane of glass.
To learn more about AWS Proton, click here.
1 ms billing for AWS Lambda
Of all the AWS Lambda announcements, this is the one that we are most excited about.
Although AWS Lambda offered an attractive price model, it relied on 100ms billing as a minimum for each execution when counting towards billing. In some cases, AWS will round up to the nearest hundred, not so great for those workloads that might not have much predictability. With this new change, AWS will bill you to the nearest millisecond with no minimum execution time.
Does that mean that there’s going to be a different way of managing lambda when it comes to billing and execution? Well, the answer is no. Starting today, you can benefit from this improvement and begin seeing some cost savings for those low latency and short-lived AWS Lambda functions.
To learn more about 1 ms billing for AWS Lambda, click here.
10 GB of RAM & 6 vCPUs for AWS Lambda
AWS has added at least a 300% increase in capacity compared to previous years. This will enable users to run more CPU intensive workloads such as Machine Learning applications or long-running analytical tasks that process data and media processing.
To learn more about 10 GB of RAM & 6 vCPUs for AWS Lambda, click here.
Container image support for AWS Lambda
For many years, developers have wanted the capacity to execute their AWS functions locally, and we’ve seen some solutions that enabled this.
With the increase in container technologies’ popularity, developers look to package their software more conveniently for working locally and running more complex applications in the Cloud.
The new Container image support for AWS Lambda allows you to leverage your current container enabled workflows and any other container solutions you may have as part of your stack, such as docker. You can do this quickly and start launching your applications using AWS Lambda.
In terms of development investments, AWS is now providing developers with base images to work with all existing runtimes and a means of testing workflow locally through the AWS Lambda Runtime Interface Emulator to make sure all applications are running as expected.
Expect AWS SAM and other popular Serverless frameworks to be compatible with this feature as well.
To learn more about Container image support for AWS Lambda, click here.