From rapidly-deployable private 5G to no-code machine learning workflows, there was a ton of stuff announced during AWS CEO Adam Selipsky’s keynote.
AWS CEO Adam Selipsky took to the stage to kick off the 10th AWS re:Invent, and he announced a lengthy list of new services available, or coming soon, from the web hosting giant.
The keynote kicking off re:Invent 2021 was lengthy, so you’d be forgiven if you didn’t watch it all, or if you simply didn’t have time to tune in. There’s a lot to cover, so let’s dive in and talk about all the new products that were announced.
New hardware and chips announced at AWS re:Invent 2021
The keynote started off with a discussion of how AWS’s EC2 instances were the first real game-change for the company, and how its next big move was its in-house Arm-based Gravitron chips it built. Today brought the announcement of the Gravitron 3 chip.
Gravitron 3 is reportedly 25% faster on general computer workloads than Gravitron 2, as well as being two times faster on scientific floating point calculations, three times faster for general machine learning workloads, and uses 60% less energy than the previous generation chip.
SEE: Hiring Kit: Database engineer (TechRepublic Premium)
The first instance to feature the Gravitron 3 is the new C7g instance for EC2, which was announced alongside it and is available in preview today.
Along with a new Gravitron 3 chip and instance type, Selipsky also announced new EC2 instances for AWS’s Trainium machine learning chip. These instances can be clustered together in the tens of thousands for intense training and ML needs, and can work along with AWS’s Inference instances, too.
Along with new chips, AWS announced the rollout of several other products best considered hardware instead of purely being a service. The first announcement, while small, is big news for users of its AWS Outpost service, which puts AWS hardware on-premise for particular use cases: There are new Outpost form factors available. Selipsky didn’t go into great detail, but described the outposts as designed “for any facility,” including rough environments like factories.
Straddling the line between hardware product and service is the next announcement: AWS Private 5G service. The reason it straddles the line is because, when purchased, it comes with all the necessary hardware to build a private 5G network. Selipsky described it as “shockingly easy, and will allow customers to set up and scale private 5G in days instead of months.”
The service is pay-as-you-go, and will arrive with all the hardware and SIM cards necessary to build out a network of the specified size for the specified number of users. Further expansion can be done in the AWS console, and there’s no per-device charge or spectrum license needed because the service operates on a shared spectrum.
New services announced at AWS re:Invent 2021
The bulk of what was announced during the keynote was new services, or new features being added to existing AWS services. Like this small one, for example: AWS Wavelength, a 5G edge computing service, is expanding into Canada in 2022, with carrier Bell Canada providing the infrastructure.
Additional smaller service announcements include serverless and on-demand analytics options for EMR, MSK, Kinesis and RedShift that can scale as needed, row and cell-level security for AWS Lake Formation and transactions for governed tables in AWS Lake Formation.
Amazon also announced a new automated digital twin making service in preview now called AWS IoT TwinMaker. TwinMaker is designed for industrial and manufacturing uses, or any case where digital twinning is a valuable part of analytics. It can be connected to IoT sensors and devices to collect data, model, and build a digital twin.
A new service called IoT FleetWise was also announced for gathering data from smart vehicle sensors for analytics. The tool can reportedly collect near real-time vehicle sensor data in order to provide reports for things like fleet health, autonomous driving performance and more.
SEE: Artificial Intelligence Ethics Policy (TechRepublic Premium)
A partnership with Goldman Sachs resulted in the next new service, called Goldman Sachs Financial Cloud for Data. Designed with capital management companies in mind, the new service brings Goldman Sachs own financial data analytics capabilities to AWS for use by customers.
The final two announcements are what have me most excited about re:Invent 2021: AWS Mainframe Modernization Service, and SageMaker Canvas.
The former, Mainframe Modernization Service, is especially welcome in the post-pandemic world, which saw mainframes and COBOL back in the news due to old, outdated mainframes being unable to handle unemployment claim loads. This brought mainframes back into public consciousness, along with the realization that everyone wants them gone, but most still rely one them.
AWS Mainframe Modernization Service may be the answer to the problem of ridiculously complex mainframe migrations. According to Selipsky, AWS MMS can cut migration time by two thirds through its near complete automation of the process. Starting with an analysis of the mainframe, MMS decides if refactoring or replatforming is the best option before providing options. It can also automate code rewriting: Selipsky specifically mentioned it can convert COBOL to Java all on its own.
The last feature is sure to excite those who want to use machine learning and its analysis capabilities, but lack the background to do so. Amazon SageMaker Canvas is a no-code module for SageMaker that uses a point-and-click interface to walk anyone through the entire process of building and using a machine learning workflow. “We democratized machine learning with SageMaker, and now we’re democratizing it even further with SageMaker Canvas,” Selipsky said.
AWS re:Invent is happening now through December 3rd in Las Vegas, and all sessions will also be available to stream online for free.