When Swami Sivasubramanian was a Ph.D. student in 2005, he landed a research engineer internship at Amazon.com Inc. He soon realized that Amazon was more than an online retailer. It was a technology company that happened to do e-commerce.

The company’s technological bias led to a white paper in 2007 that outlined the concept for a new database architecture called DynamoDB, which could run high-performance applications at any scale. The project was born out of a pressing need for a new solution after a hectic holiday sales season nearly crashed Amazon’s extensive computing capacity.

“When we had a major disruption during Q4, we asked ourselves this question: ‘Why are we using a relational database for these things when they really didn’t need the data model complexity of a relational database?’” recalled Sivasubramanian (pictured). “That actually let us start reimagining what a database for our scale could look like. It led to Dynamo.”

Sivasubramanian spoke with John Furrier, host of theCUBE, SiliconANGLE Media’s livestreaming studio, in an exclusive interview. They discussed the factors behind DynamoDB’s development, how Amazon’s customer-oriented philosophy guided the company’s approach and the role of cloud in driving machine learning.

Pivotal moment for cloud

January marked the 10-year anniversary of DynamoDB’s initial release. The database supports key-value cloud services with features such as auto-scaling, in-memory caching, and backup/restore for internet-scale applications. In addition to supporting Amazon, DynamoDB also powers customers such as Zoom, Disney, Dropbox, Samsung, Capital One and Netflix.

“One of the big revelations we had was there was not a single database that was going to meet customer needs as the diversity of workloads on the internet was growing,” Sivasubramanian said. “This was a key pivotal moment, because now with cloud, applications can scale way more instantly than before.”

Part of DynamoDB’s development relied on a principle within Amazon of always working back from the customer. It’s a philosophy that has driven much of the company’s innovation for nearly three decades, although DynamoDB followed a slightly different path.

“Many times, when we work backwards from customers, the literal way to translate it is do what customers are asking for, which is true nine out of 10 times,” Sivasubramanian said. “But there is one out of 10 times where you’ve got to read between the lines on what they are asking. They might say they wish their horse carriage could go faster, but they’re not going to tell you they need a car. We call it innovating on behalf of customers, and that is the kind of mantra we had when we were thinking about concepts like DynamoDB.”

Machine learning for all

DynamoDB represents another chapter in the ongoing saga of technological change driven by the cloud. Sivasubramanian’s paper, published as part of a collaborative effort with Amazon Chief Technology Officer Werner Vogels and seven others, appeared one year after the company launched a new cloud services business known as Amazon Web Services Inc.

Sivasubramanian would go on to hold a series of positions with increasingly greater responsibility within the company, culminating in his current role overseeing machine learning services within AWS. Much as the cloud influenced his approach to creating DynamoDB 15 years ago, it’s shaping the technologist’s thinking today when it comes to machine learning as well.

“Cloud has basically enabled limitless compute and limitless storage, which are the factors that were holding back machine learning technology,” Sivasubramanian said. “I realize that now we have a unique opportunity to bring machine learning to everybody. Machine learning will be one of the most disruptive technologies we will encounter in our generation.”

Here’s the complete video interview, one of many CUBE Conversations from SiliconANGLE and theCUBE:

Photo: SiliconANGLE