Home Artificial Intelligence The Future of Artificial Intelligence, Cloud Computing, and Digital Twins

The Future of Artificial Intelligence, Cloud Computing, and Digital Twins

by Bernard Marr
0 comment

Befitting its status as the first cloud provider to offer cloud compute, storage, and database services – and the transformational effect this has had, in IT services and far beyond – the theme of Selipsky’s talk was “pathfinders.” As well as highlighting the company’s success stories over the years he covered a selection of inspirational individuals from history who themselves fit that mold. These included Florence Nightingale – herself a “data geek” – as Selipsky put it – who harnessed statistics and visualizations to transform the field of medical care, as well as Hank Luisetti – “the most disruptive basketball player of all time,” and Roscoe Brown.

Brown was a decorated war hero who flew over 60 combat missions during World War 2 before being turned down for a civil aviation job due to the color of his skin. He went on to dedicate his life to the civil rights struggle and helping young African Americans overcome prejudice.

As was expected, Selipsky used his keynote to announce the launch of a number of new AWS products and services. Here’s a rundown of some of the most exciting developments, as well as a look-ahead to what else we can expect over the next few days of AWS 2021:

Graviton 3

Selipsky used his keynote on Tuesday morning to announce the latest version of AWS’s Graviton processor. The chip is designed to speed up machine learning operations by up to three times and is two times faster at floating-point operations. As with previous generations of the Graviton chip, it is designed around ARM architecture. Importantly, the chip is also 60% more power-efficient than previous models, allowing it to play a part as businesses move towards more sustainable and eco-friendly models of computing and IT infrastructure.

Selipsky also announced that Graviton 3 would be made available to AWS customers very soon through a new instance, called C7G, and designed specifically for intensive and high-performance compute (HPC) workloads in the Amazon cloud.

TRN1 – Tranium instances

Selipsky announced another new instance developed specifically for AWS customers looking to run highly intensive compute workloads, including machine learning. The TRN1 EC2 instances are designed for running training operations, including image recognition, natural language processing, and fraud detection. It accommodates network speeds of up to 800GB per second, making it suitable for the largest scale enterprise use cases. The instances are designed to be networkable into “ultra clusters” of tens of thousands of instances, able to crunch through petabyte-scale data workloads.

Amazon Sagemaker Canvas

Contributing to the ongoing drive to put machine learning and analytics abilities in the hands of as many people as possible – the “democratization of data science” – Selipsky announced the launch of Canvas for Amazon’s machine learning service, SageMaker.

Canvas is a “no-code” solution that uses visual tools to allow anyone with no formal background in data science to start analyzing, interrogating, and querying data in an intuitive way, automating tasks such as data cleansing and transformation while making it simple to train models and generate insights from the information. It is focused on business use cases, such as fraud detection and process optimization, rather than academic uses. With the launch of tools like SageMaker Canvas, it’s great to see Amazon further committing itself to the task of enabling as many people as possible to take advantage of the opportunities of AI and ML.

Selipsky said, “We built Amazon Sagemaker as a way to democratize machine learning … so we asked ourselves whether there’s something we can do to further democratize ML by helping an entirely new group of users with no ML experience – no data science experience – who are not even developers, to be able to do machine learning. Today I’m excited to announce a new capability of SageMaker, to enable business users and analysts to generate highly accurate machine learning predictions … with no coding required.”

AWS IoT TwinMaker

Digital twins are essentially virtual copies of real-world objects, systems or processes – anything from a simple machine to an entire city. IoT TwinMaker makes it faster and easier for developers to create digital twins of real-world systems like buildings, factories, industrial equipment, and production lines. It can be used to run simulations in the virtual domain in order to gather data on how the subject might perform in the real world – at a fraction of the time and expense.

IoT TwinMaker connects to real-world data sources and automatically generates “knowledge graphs” that map out the relationships between the data sources and operational characteristics of whatever is being twinned. All of this happens via 3D models that update in real-time as the data changes. AWS has established partnerships with Siemens, Carrier and Accenture to integrate the service with industrial IoT services.

AWS Private 5G

One of the most exciting announcements was the pilot launch of AWS’s Private 5G service, which allows any organization to build and host their own 5G networks on a pay-as-you-go basis. Restricted for now to US customers, this has the potential to boost innovation in the 5G space by lowering the barrier to entry, as start-ups will no longer need to invest in expensive, bespoke 5G infrastructure such as masts and towers in order to run 5G networks for their users and applications. 5G has huge speed, connectivity, and security advantages compared to previous generations of mobile data network and is expected to enable many new services based on new types of data, rather than simply faster transmission of existing types of data. Selipsky said that this will allow businesses to plan, deploy and scale 5G networks and applications in “days instead of months.”

U.S.-based connectivity company DISH Network Corporation has already started using the private network. Stephen Bye, Chief Commercial Officer, DISH, said, “Our ability to support dedicated, private 5G enterprise networks allows us to give customers the scale, resilience and security needed to support a wide variety of devices and services, unlocking the potential of Industry 4.0.”

NASDAQ’s Market Replay matching engine

Selipsky was joined on stage by Adena Friedman, NASDAQ president and CEO, to announce the migration of its US options markets into the Amazon cloud. This relates to the company’s Market Replay service, launched in 2008 on AWS, which allows customers to review an entire day of market activity.

“We will work with AWS to bring our first US options market to the AWS cloud in 2022”, Friedman told the audience. The plan is to then follow this up by moving more markets into the cloud in the near future. The technology will also be made available as a service to NASDAQ’s customers, who make use of market technology themselves for trading and clearing operations. The plan is to make markets “safer, stronger and more accessible.”

AWS Mainframe Modernisation

This is a new service designed to let customers more efficiently migrate legacy mainframe workloads onto the AWS cloud environment. It allows applications that were written in an era before the standards of web and cloud software were established, porting code into Java that will run on AWS servers while automating the process of re-platforming the workload. “Many workloads today still run in your datacenter because you’ve deployed so many workloads there,” Selipsky said.

“AWS pioneered cloud infrastructure services … but we’ve also been building bridges back to your data centers… to make it possible for you to use the same familiar software and tools that you know and love in your data centers, seamlessly in AWS.”

Selipsky also announced the availability of smaller form-factors for AWS’s Outpost hybrid cloud infrastructure, which allows AWS functionality to be deployed in any on-premise environment, including “rugged edge” environments like oil rigs and agricultural fields.

Goldman Sachs Financial Cloud for Data with AWS

Selipsky also announced a new service developed in partnership with Goldman Sachs, designed to empower financial services companies through data, ML, and analytics. It involves Goldman Sachs making its data and tools available to hedge funds and asset managers through an Amazon cloud service and marks a step towards becoming a fully-fledged technology service provider for the grandfather of US investment banks.

The service is designed to allow financial services companies to focus on trading and investment without having to worry about building and maintaining data infrastructure and data science operations such as data cleansing and validation.

AWS Fleetwise

This is a new service designed to help businesses involved with smart, autonomous, and connected cars to manage data flowing between the vehicles, as well as the edge devices and data centers involved in their operation. It provides a platform for tools enabling in-car sensors, autonomous driving, and remote diagnostics and maintenance; all managed through an AWS interface.

Selipsky told the audience during his AWS 2021keynote, “The cloud is fundamentally changing this industry … including how vehicles are designed and manufactured, the features they offer and how we drive, and it’s all happening at Formula 1 speed, as manufacturers … are designing vehicles that are infused with software and connected by sensors, and systems generating unheard of amounts of data.

“We will see cars that are intelligent and autonomous, energy-efficient and inexpensive to maintain. However, the sensors on one car can generate up to two terabytes of data every hour … it’s easy to understand why it’s necessary to build custom data collection systems … but building these systems is difficult and time-consuming.”

You may also like