AWS Delivers More AI, Analytics, Partnerships And A Big Rig For Moving Big Data

Amazon Web Services (AWS) showcased several artificial intelligence and data services Wednesday for partners to use in building applications that evaluate photos, talk to users and gain insights from their data.

The new AI, database and analytics solutions add to a broad portfolio of tools for ingesting and evaluating data, AWS CEO Andy Jassy told attendees of the AWS re:Invent conference in Las Vegas.

The cloud leader is "making petabyte-scale analytics accessible to companies large and small," Jassy said.

[Related: AWS Vastly Diversifies Its Portfolio Of VMs And Aims To Attract More Industries To Cloud Computing]

A trio of new services will enable partners to build intelligent interfaces.

"Customers are anxious to understand their data better" and AI promises to do that, Jassy said.

Amazon Rekognition can detect objects in an image, identify people, match faces, even perform a sentiment analysis with facial recognition, determining if the person in the photo is smiling or frowning or wearing glasses.

Amazon Polly is a text-to-speech deep learning service. Polly ingests written text, and returns an MP3 stream in natural sounding language—with 47 different voices in 27 different languages to choose from.

And Amazon Lex—because it's what's inside of Alexa, the AI engine that powers the Echo—offers speech recognition and natural language comprehension.

Matt Wood, general manager of data sciences at AWS, told partners the three services would enable them to build "multi-step, sophisticated, conversational interfaces."

More AI features are coming, Jassy said, part of Amazon's initiative to make it easier for everyday developers to do machine learning—a technology that is a deep part of Amazon's heritage. Thousands of Amazon employees focus solely on AI, he said.

"Some of this stuff is for the tech, tech, tech companies," said Jeff Aden, executive vice president at 2nd Watch, one of AWS' largest partners. But enterprises will probably adopt those technologies in the coming years after exploring the use cases.

While not many 2nd Watch customers are immediately clamoring for those capabilities, they are exciting because they so directly showcase the innovation happening at AWS.

Jassy said some customers want to run simple analytics, processing logs or click streams, and make their queries directly from S3 storage.

Amazon Athena is an interactive query service that makes that possible using standard SQL.

"There's no infrastructure to spin up and manage, no clusters to spin up and manage," Jassy said of Athena. Results come fast, and you pay for only the queries you run.

Jassy also introduced PostgreSQL on Aurora, delivering database compatibility that many partners and customers had been requesting.

Jassy took a low-key swipe at a competitor during his keynote, mentioning the "hand-waving and bombast" of "old-guard leaders" trying to sell their products as a picture of Oracle founder Larry Ellison briefly popped up on the corner of the screen. Ellison assailed AWS at his company's conference in October.

Customers don't need to suffer such theatrics, Jassy said, because the cloud gives them enough information to make a buying decision—they can benchmark and test services before committing.

Another important announcement from Jassy's keynote: cloud-based ERP vendor Workday has decided that AWS is going to become their preferred public cloud provider for production workloads.

Workday co-founder and CEO Aneel Bhusri said the company is moving all its HR and accounting applications to Amazon's cloud.

Workday will roll out those services progressively, starting in 2017 with Canada, then in the US and other parts of the world

Jassy also welcomed VMware CEO Pat Gelsinger onto the stage to further discuss the partnership between the two companies announced about six weeks ago.

The VMware partnership adds to a "broad number of tools" that enable hybrid cloud operations, he said.

"We've seen in the last several months as many more large enterprise are exploring substantial shifts to the cloud, they say you are forcing us to make a binary choice," Jassy said.

The choice was indeed unfavorable to many customers "because nearly every enterprise in the world is virtualized on VMware," Jassy said, and AWS and VMware previously wouldn't work together. The upcoming VMware Cloud For AWS offering will solve that dilemma.

The most surprising product introduction in the keynote came as a white 18-wheeler drove up onto the stage.

Last year, AWS introduced Snowball, a physical case that holds 50-Terabytes of storage for migrating data to the cloud. They upgraded that product this year with Snowball Edge, which adds compute capabilities to the box and doubles the storage.

But Jassy said some customers are already considering migrations involving exabytes of data, and that would take thousands of such boxes.

The data-hauling big rig, called Snowmobile, addresses that problem.

A data transfer that would take roughly 26 years with a 10 Gbps dedicated connection can be accomplished in months with 10 Snowmobiles driving up to a customer data center, connecting with a fiber cable, and drives all that data back to AWS, Jassy said.

Do customers need to move that much information?

"There are companies moving lots and lots of data," 2 nd Watch's Aden said.