Amazon Web Services today revealed several new cloud services ranging from new databases to new AI and machine learning tools, some with particular importance for Australian customers, according to the company.
Over 100,000 independent enterprises in Australia, ranging from startups to the big banks and government agencies, are running at least one of AWS’s services – now 175 total following today’s announcements.
AWS managing director, ANZ, Paul Migliorini says Australia is a mature market for public cloud and more organisations are migrating as they realise the technology can solve business problems.
“I think [we are] giving people the tools to go prosecute business problems and then helping them change the culture of their organisation, which is what is happening a lot of the time,” Migliorini told Which-50 at the AWS re:Invent conference in Las Vegas today.
The local MD said around 90 per cent of the services AWS builds come directly from customer feedback with the remaining 10 per cent being “iterating on their behalf”.
“When you’ve got a platform now that has 175 services ranging from compute all the way through to AI you’re seeing a lot of iteration on a lot of levels … It’s a platform that’s enabling that innovation.”
Throughout a three hour keynote today, AWS CEO Andy Jassy announced a host of new and updated services, including a new machine learning interface, the release of an AWS infrastructure edge unit, and a new fraud detection product, among others.
AWS Outposts are fully managed and configurable racks of AWS hardware that allow customers to run their workloads on premises in the same way they would in the cloud. Announced last year, AWS Outposts were made generally available today.
There is also a VMWare version of the Outpost which runs the virtualisation company’s architecture but it will not be available until some time in 2020.
Migliorini told Which-50 his Australian resource customers are particularly excited about Outpost which will bring some of AWS’s services to the edge in remote environments.
“We’re seeing lots and lots of customers, whether it’s kind of agri businesses or mining or gas businesses or government agencies, looking for those kind of single digit millisecond latency [times] and an ability to deploy an equivalent version of the AWS cloud in their own data centre.”
AWS also made a significant addition to its main machine learning tool, known as SageMaker. Several AWS customers Which-50 spoke with said the SageMaker updates were the most important updates of the event.
Amazon SageMaker Studio adds an integrated development environment functionality to SageMaker, allowing developers to collect models, run debuggers, and monitor the development and deployment of ML models.
Migliorini said there are several thousand Australian customers using SageMaker already. According to the ANZ MD, local customers are using SageMaker, which aims to democratise machine learning technology, because it helps fill Australia’s current machine learning skills gap.
Machine learning is also powering a new fraud detection service AWS will now begin offering. According to the company it uses some of the models Amazon uses itself to detect fraudulent activity on its ecommerce business.
Amazon Fraud Detector is a fully managed service with AWS handling the machine learning. Customers submit historical data and identify transactions which were fraudulent. AWS then trains a model to detect fraud and deploys it via a a private API, allowing customers to run it in real time.
Amazon says even if customers’ datasets do not need to be as large as is usually required for machine learning models because it is adding some of its own historical data and models to help detect fraud.
The author traveled to AWS re:Invent as a guest of Amazon.