Highlights from Andy Jassy’s AWS re:Invent 2020 Keynote

Ross Lawrie

two people looking at computer screen, with one person pointing to the screen


AWS re:Invent 2020 is off to a huge start, with Andy Jassy presenting a tsunami of new services and features during his opening keynote. From databases to machine learning, from cloud compute to hybrid solutions, from containers to disk storage, there’s far too much to write about in a single entry, so here are just a few of the highlights.


Babelfish for Amazon Aurora PostgreSQL

One of the biggest announcements is massively important to customers making the migration from on-premises facilities to the cloud, and their challenges around Microsoft SQL Server in particular. This often represents a large cost to operate due to licensing, but the move to an alternative engine like PostgreSQL is daunting due to the work required to update applications and queries. With Babelfish for Aurora PostgreSQL, customers can enable a SQL Server compatible endpoint to their Aurora PostgreSQL cluster, which can handle T-SQL and TDS communications from existing applications, with no updates necessary. This will allow you to quickly remove or reduce expensive licenses without the headaches of building new schema or changing all your queries. As an added bonus, AWS will make Babelfish open source in 2021 so it can be validated and improved by the community. Babelfish is available in preview right now.


Amazon ECS Anywhere and Amazon EKS Anywhere

With containers being such a key piece of many environments these days, it seemed likely that there’d be at least one announcement here, and there’s a big one, or two. Amazon ECS Anywhere and Amazon EKS Anywhere allow organizations to deploy EKS and ECS tasks to any environment, whether that’s the usual AWS infrastructure, or your on-premise infrastructure. The control pane continues to deploy within AWS, but from there you can use the regular service API to deploy and manage containers in either environment. This will be extremely useful in allowing organizations to use existing hardware while still driving towards their goals of moving to the cloud. Additionally, this can power low latency interactions within local data centers for such requirements. Both services will be available next year.


Container Image Support for AWS Lambda and AWS Proton

Next up is Lambda. A frequent challenge for customers using Lambda is managing the deployment of their functions, as well as the management of growing microservice architectures powered by Lambda. With that in mind, two announcements today aim to alleviate that: Container Image Support for Lambda and AWS Proton. Container Image Support enables packaging Lambda functions as container images, allowing teams to leverage all their existing container build and deployment pipelines for their Lambda functions as well. AWS Proton is a fully managed deployment service for container and serverless applications, aimed at making the management of sprawling microservices much easier. Application stacks are defined, containing everything from the infrastructure to the deployment pipeline, greatly simplifying the management of the CI/CD flow. Proton also allows for inspection and monitoring of the deployment processes, providing a clear trail on who deployed what, and when. Both are generally available today.


Io2 Block Express

With a big focus on solving the implementation of SANs within the cloud, io2 Block Express was announced today, allowing for the deployment of massive and highly performant block volumes – up to 64 TiB, 256,000 IOPS and 4000 MBps of throughput, which is up to 4x the performance of the regular io2 volumes. The service is available in preview currently, and AWS is planning the release of additional features next year such as Multi-Attach, Elastic Volumes, I/O fencing and Snapshot Restore. io2 Block Express EBS volumes are currently in preview.


Amazon Connect

Another highlight was all the enhancements made to the call center service, Amazon Connect, which allows for the easy creation and management of a call center. The enhancements look to leverage machine learning to improve customer and agent interactions. Amazon Connect Wisdom uses ML to detect the content of calls and places relevant information in front of the agents to help customers quicker and more accurately. Amazon Connect Customer Profiles are able to unify customer information from multiple sources to make insights about the customer available to the agents. Negative customer experience issues can be detected in real-time during a call by Amazon Connect Real-Time Contact Lens, which can be used to escalate the calls to supervisors while providing them with transcriptions on the call so far. Amazon Connect Tasks unifies the management of all the tasks associated with the call center and allows easy routing through existing tools. Last, but not least, Amazon Connect Voice ID uses ML to build voice signatures for callers, allowing for identification through natural conversation and removing the need for authorization questions by agents based on the certainty of the model.


This only hits on some of the bigger announcements during the keynote, and doesn’t dive into high performance gp3 EBS volumes, Amazon DevOps Guru which provides machine learning derived insights into infrastructure and configuration, SageMaker Pipelines that provides CI/CD for ML workflows, smaller AWS Outpost devices (1U and 2U forms) and more!

Andy Jassy’s keynote is an amazing kick-off to this year’s event, clearly showing AWS’ drive to provide customer-driven services that help inspire invention and accelerate cloud adoption. With three weeks of keynotes and sessions for re:Invent this year, there are even more big announcements in store that we’ll be continuing to highlight here.


Get Help at any Stage of Your Cloud Journey