Improvements to its SageMaker machine-learning tool and new scaling options for cloud computing customers were among the new features rolled out by Amazon Web Services this week, days before it will probably announce dozens of new features and products at its annual re:Invent conference in Las Vegas.
Apparently a few things had to be cut from the marathon keynote sessions scheduled for next week by CEO Andy Jassy, Amazon CTO Werner Vogels, and vice president of infrastructure Peter DeSantis, prompting AWS to reveal several new enhancements to its services a week early. The event is expected to attract over 40,000 people for a week of technical sessions, announcements, and networking with basically the entire cloud computing industry sharing the hallways of the Venetian.
AWS announced Wednesday that SageMaker, a service that helps developers get up and running with machine learning models, now comes with better search and workflow management capabilities. Customers will also be able to select new algorithms, such as one that detects suspicious IP addresses, and satisfy compliance requirements around their machine-learning practices.
The company’s Snowball Edge devices are also getting an upgrade, with a new option to purchase a version with an on-board GPU (graphics processing unit) that can handle machine-learning tasks or video analysis. The original Snowball Edge device now comes with a more powerful processor, AWS evangelist Jeff Barr wrote in a blog post.
And one of its most basic services — the EC2 compute service — can now predict when a customer’s application will see a spike in demand for computing resources based on past activity. This could help customers improve the performance of their applications and also get a better handle on how much they need to spend with AWS.
We’ll have coverage of re:Invent 2018 starting on Monday from Las Vegas. Please send backup batteries.