The Future of Amazon Web Services: 5 Things to Know From AWS Re:Invent 2021

Unlike last year when it was entirely virtual, Amazon Web Services' (AWS) re:Invent show was held in meatspace this year. However, a virtual registration was also available (and you can still register here). If you're an IT pro with any AWS infrastructure or service in your portfolio, you should because the company has bombarded customers with a mudslide of news.

AWS has updated its data analytics services, announced new hardware specs for its virtual machines, and upped its game in enterprise robotics. It's given customers a faster way to dump their mainframes, and it's also shown a preview of its upcoming managed 5G service. Here's our quick take on the top AWS announcements so far and why they're essential.

1. Data Analytics Gets Beefed Up With AI

The What: AWS has added some muscle to its previous machine learning (ML) and artificial intelligence (AI) tools as applied to its suite of analytics services. The new stuff is aimed at partners who now have access to AWS' AI for Data Analytics (AIDA) solutions that'll let them increase their predictive analytics sophistication while keeping the tools simple enough for non-data scientists to use. AWS touted its Cognizant, Interworks, Provectus, and Quantiphi partners as being early adopters here. It also announced much more granular controls for customers' data lakes as well as a new serverless approach to most of its data services.

Why You Care: Predictive analytics lets your execs make better decisions on future developments. So these algorithms can be beneficial if you're analyzing sales trends, customer behavior, and the future costs of anything for which you've got some data history. And you can store the data and run these analytics tools on a cloud service, which will seriously cut your operational costs.

Data lakes let you store big gobs of data without all the traditional data warehouse structure chores. This will also up-level analytics access to people without deep data science backgrounds. The new tools will let you add and control data access policies right down to the individual row and cell level and do it on an easy policy basis.

According to AWS, customers of its Amazon EMR, Amazon Kinesis, Amazon MSK, and Amazon Redshift analytics platforms have long been asking for them to go serverless and on-demand. This year, AWS says it's delivered on that ask. That's important since if you're a Kinesis customer, for example, you're running streaming data through the service. Because it's now fully managed and on-demand, it can scale automatically, covering both resources and cost. Or if you've got a data warehouse in Redshift, AWS says access should be easier and faster, plus you're only getting charged when the warehouse is active.  

2. Upgraded Hardware Specs

The What: To cut costs across services and virtual infrastructure (which should be welcome news to any AWS customer), the company announced new, higher-performance hardware for the underlying infrastructure that powers its cloud. AWS says the new machines are running the third generation of AMD’s EPYC processors, the M6a. The company also announced VM options around the new Graviton ARM processor, which it's coupling with Nvidia T4G Tensor GPUs.

Why You Care: Hey, it's more VM muscle; who doesn't like that? But specific benefits beyond just processing oomph is reduced overall costs, at least according to AWS. The company believes more power means fewer VMs necessary to particular workloads, which should drop customer costs. AWS claims the M6a option will juice up to a 35% performance bump over what it's running now, based on AMD M5a CPUs. For virtual machine (VM) users, this will translate into new software infrastructure options scaling from two to 192 virtual CPUs with RAM sizes to match. The Graviton-to-Nvidia match is meant for customers serving up games and says the combination should drop the costs for those workloads by up to 30%. And it should also make data scientists happier if they happen to be working on resource-sucking ML algorithms. Will you actually see those cost benefits? Sure, why not?

The What: If your company uses robotics to drive any kind of Internet of Things (IoT) workload, AWS wants you to develop those resources in its cloud. To help, it's just launched a public preview of the new AWS IoT RoboRunner. And if you're thinking about starting a robotics business, there's a new accelerator program, too.

Why You Care: Besides being hard to say, AWS IoT RoboRunner is exactly what it sounds like: a cloud-managed robot management system. If you're watching a few dozen or even hundreds of burgeoning R2D2s, this tool lets you see them all through a single pane of glass with customizable health and management features.

The AWS Robotics Startup Accelerator is meant for entrepreneurs and their technical leads to understand both the robotics tech and business landscape. It's a four-week mentorship-style program that will be available globally, and it's being delivered via an alliance AWS now has with MassRobotics, a non-profit innovation hub specializing in robotics technology.

4. A Peek at 5G, AWS Style

The What: Re:Invent 2021 also gave us a sneak peek at AWS Private 5G. The service provides services to help customers deploy their networks based on or extended to 5G technology.

Why You Care: 5G is new and therefore full of deployment and implementation nightmares. According to AWS CEO Adam Selipsky, the service will let you spin up a private 5G mobile network in just a few days. Sure, that claim is probably overly optimistic hype, but if it's even only half true, that's still a massive dent in the average 5G implementation cycle. Customers will be able to describe their capacity and location requirements, and AWS Private 5G provides everything you need right down to the SIM cards.

5. Get Off Your Mainframe Now

The What: If you've still got a basement data center stuffed with aging mainframe technology running some proprietary or legacy software, the AWS Mainframe Modernization service is there to help you move to the cloud.

Why You Care: If you don't have a mainframe, you don't care. If you do, then this isn't help with modernization; it's help with migration. The service offers a set of assessment and analysis tools to help customers understand what they're running and form a plan to help Amazon slurp your workloads into AWS. Once the plan is there, the service has a list of new developer tools, including features around software testing and deployment. The potential downside is that the AWS service is aimed at partners to resell the expertise to their customers. That'll mean added costs, but if you're spinning your wheels on a mainframe migration that's getting more necessary by the day, a partner contract is still much better than nothing.

The above are the five announcement buckets that jumped out at PCMag. There are many more, particularly the new partner programs, which is largely the intended audience for re:Invent. These programs will help partners with everything from training consultants and technicians to specialization tools AWS calls “partner paths.” These are business tools and educational resources that'll help partners deepen their expertise in areas as diverse as analytics, migration, and third-party service deployment.

You can go deeper into any of these announcements by registering for the event, which will be open through the morning of Dec. 3. However, most of the content should also be available after the show on-demand.