AWS re:Invent 2021 kicks off with an announcement-filled keynote
7 mins read

AWS re:Invent 2021 kicks off with an announcement-filled keynote

AWS re:Invent 2021 kicks off with an announcement-filled keynote

From rapidly-deployable personal 5G to no-code machine studying workflows, there was a ton of stuff introduced throughout AWS CEO Adam Selipsky’s keynote.


AWS CEO Adam Selipsky kicks of re:Invent 2021.

AWS CEO Adam Selipsky took to the stage to kick off the tenth AWS re:Invent, and he introduced a prolonged record of latest providers out there, or coming quickly, from the internet hosting large. 

The keynote kicking off re:Invent 2021 was prolonged, so you would be forgiven in case you did not watch all of it, or in case you merely did not have time to tune in. There’s lots to cowl, so let’s dive in and discuss all the brand new merchandise that have been introduced. 

New {hardware} and chips introduced at AWS re:Invent 2021

The keynote began off with a dialogue of how AWS’s EC2 cases have been the primary actual game-change for the corporate, and the way its subsequent large transfer was its in-house Arm-based Gravitron chips it constructed. At present introduced the announcement of the Gravitron 3 chip.

Gravitron 3 is reportedly 25% quicker on basic laptop workloads than Gravitron 2, in addition to being two instances quicker on scientific floating level calculations, thrice quicker for basic machine studying workloads, and makes use of 60% much less vitality than the earlier era chip.

SEE: Hiring Package: Database engineer (TechRepublic Premium)

The primary occasion to function the Gravitron 3 is the brand new C7g occasion for EC2, which was introduced alongside it and is on the market in preview immediately.

Together with a brand new Gravitron 3 chip and occasion sort, Selipsky additionally introduced new EC2 cases for AWS’s Trainium machine studying chip. These cases will be clustered collectively within the tens of hundreds for intense coaching and ML wants, and may work together with AWS’s Inference cases, too. 

Together with new chips, AWS introduced the rollout of a number of different merchandise finest thought-about {hardware} as an alternative of purely being a service. The primary announcement, whereas small, is large information for customers of its AWS Outpost service, which places AWS {hardware} on-premise for specific use instances: There are new Outpost type elements out there. Selipsky did not go into nice element, however described the outposts as designed “for any facility,” together with tough environments like factories.

Straddling the road between {hardware} product and repair is the subsequent announcement: AWS Non-public 5G service. The explanation it straddles the road is as a result of, when bought, it comes with all the required {hardware} to construct a personal 5G community. Selipsky described it as “shockingly simple, and can enable prospects to arrange and scale personal 5G in days as an alternative of months.” 

The service is pay-as-you-go, and can arrive with all of the {hardware} and SIM playing cards mandatory to construct out a community of the required dimension for the required variety of customers. Additional growth will be finished within the AWS console, and there is no per-device cost or spectrum license wanted as a result of the service operates on a shared spectrum. 

New providers introduced at AWS re:Invent 2021

The majority of what was introduced in the course of the keynote was new providers, or new options being added to current AWS providers. Like this small one, for instance: AWS Wavelength, a 5G edge computing service, is increasing into Canada in 2022, with service Bell Canada offering the infrastructure.

Extra smaller service bulletins embrace serverless and on-demand analytics choices for EMR, MSK, Kinesis and RedShift that may scale as wanted, row and cell-level safety for AWS Lake Formation and transactions for ruled tables in AWS Lake Formation. 

Amazon additionally introduced a brand new automated digital twin making service in preview now referred to as AWS IoT TwinMaker. TwinMaker is designed for industrial and manufacturing makes use of, or any case the place digital twinning is a beneficial a part of analytics. It may be linked to IoT sensors and gadgets to gather knowledge, mannequin, and construct a digital twin.

A brand new service referred to as IoT FleetWise was additionally introduced for gathering knowledge from sensible automobile sensors for analytics. The device can reportedly gather close to real-time automobile sensor knowledge with a purpose to present studies for issues like fleet well being, autonomous driving efficiency and extra. 

SEE: Synthetic Intelligence Ethics Coverage (TechRepublic Premium)

A partnership with Goldman Sachs resulted within the subsequent new service, referred to as Goldman Sachs Monetary Cloud for Information. Designed with capital administration corporations in thoughts, the brand new service brings Goldman Sachs personal monetary knowledge analytics capabilities to AWS to be used by prospects. 

The ultimate two bulletins are what have me most enthusiastic about re:Invent 2021: AWS Mainframe Modernization Service, and SageMaker Canvas. 

The previous, Mainframe Modernization Service, is very welcome within the post-pandemic world, which noticed mainframes and COBOL again within the information as a result of outdated, outdated mainframes being unable to deal with unemployment declare masses. This introduced mainframes again into public consciousness, together with the belief that everybody desires them gone, however most nonetheless rely one them.

AWS Mainframe Modernization Service often is the reply to the issue of ridiculously complicated mainframe migrations. In response to Selipsky, AWS MMS can minimize migration time by two thirds by its close to full automation of the method. Beginning with an evaluation of the mainframe, MMS decides if refactoring or replatforming is the best choice earlier than offering choices. It could possibly additionally automate code rewriting: Selipsky particularly talked about it might convert COBOL to Java all by itself.

The final function is certain to excite those that wish to use machine studying and its evaluation capabilities, however lack the background to take action. Amazon SageMaker Canvas is a no-code module for SageMaker that makes use of a point-and-click interface to stroll anybody by the whole strategy of constructing and utilizing a machine studying workflow. “We democratized machine studying with SageMaker, and now we’re democratizing it even additional with SageMaker Canvas,” Selipsky stated. 

AWS re:Invent is going on now by December third in Las Vegas, and all classes may even be out there to stream on-line at no cost.

Additionally see

Leave a Reply

Your email address will not be published. Required fields are marked *