Amazon Web Services Follows Microsoft's Lead With Cloud Machine Learning
Amazon Web Services has benefited from a giant first-mover advantage in the cloud market, and it unveiled several new cloud services Thursday that could make it more difficult for rivals to catch up.
However, AWS is playing catch-up to Microsoft Azure with its new Amazon Machine Learning service, which it debuted at its annual AWS Summit in San Francisco. The service, available now, is similar to a service that Microsoft began selling in February.
Amazon is packaging the machine learning it uses in its retail business and making the complex technology available to developers in an easily usable form, Andy Jassy, senior vice president of AWS, said in a keynote at the event.
Machine learning is a term that describes how computers can improve their functioning over time by ingesting and analyzing data. Predictive analytics is a form of machine learning that crunches historical data to predict what will happen in the future.
Amazon uses machine learning on the retail side to figure out what users want to search for, detect counterfeit goods sold by third parties, gauge demand for products and support customers, Jassy said.
Dr. Matt Wood, general manager of data science at AWS, said in the keynote that Amazon Machine Learning lets developers build predictive models that automatically find patterns, and use them "to make confident predictions."
Machine learning is one area of the cloud market where Microsoft beat AWS to the punch. Microsoft poached a former Amazon machine-learning executive, Joseph Sirosh, in 2013 to help lead its push into this nascent market segment.
Meanwhile, AWS also debuted a new service at the event called Elastic File System (EFS), which works with Amazon EC2 instances and fills what has been a "missing piece in the cloud," Jassy told attendees.
EFS automatically replicates file systems to multiple AWS availability zones so that EC2 instances in the same regions can use them, said Jassy. The AWS EFS service also supports the Network File System version 4 (NFSv4) protocol.
Luis Benavides, founder and CEO of Day1 Solutions, an AWS partner in McLean, Va., sees EFS as another example of how AWS listens closely to what its enterprise customers are asking for. "Instead of customers having to use a Windows file server or third-party tools, now AWS is doing it themselves," he told CRN.
Christian Puopolo, a systems specialist at Axispoint, a New York-based AWS partner, said he sees the new service as "an auto-scaling secure cloud SAN for no money down."
"AWS Elastic File System has the potential to effect positive change across all AWS services," Puopolo said in an email. "It can greatly reduce costs by eliminating the footprint of a deployment, while increasing the flexibility, speed, and responsiveness."
AWS is also beefing up functionality for Workspaces, its Desktop-as-a-Service offering, with a new online marketplace for popular apps that customers run on virtual desktops.
Called AWS Marketplace for Desktop Apps, it lets customers purchase apps and have them provisioned to run on AWS Workspace virtual desktops. And instead of paying upfront for software licenses, customers pay for the apps on a monthly subscription basis, Jassy said at the event.
AWS also unveiled Workspaces Application Manager, which manages virtual apps, runs them as if they're installed natively, and automatically manages patches, backups and upgrades.
AWS wasn't first to launch a cloud machine learning service, but its other new additions show that it's focused on continuing to put pressure on enterprise vendors.
PUBLISHED APRIL 9, 2015