A self-driving car seems relatively straightforward in concept, even if the mechanics behind it are extraordinarily complex. But what if that idea can be applied elsewhere, specifically to data centers?
That's the work that Balaji Prabhakar is doing. He is a computer science professor and faculty member at Stanford University, and he spoke with Nasdaq's Tom Fay, who leads the Enterprise Architecture and Systems Engineering team, on the most recent episode of Tomorrow's Capital.
Prabhakar said, "My area of research has been computer networks, and that's what I've primarily been working on, [seeing how] self-driving networks are related to computer networks, and to borrow something from the world of self-driving cars. The idea is that if we take a car just off the street, and add lidar sensors to it, and control the steering and gas pedal, then we get a self-driving car. It can move around. It knows how to navigate and avoid obstacles.
"Similarly, managing a data center or what is basically the networks that are underlying Cloud computing infrastructures is actually a complex task. You have to schedule many different resources. The workload that is hitting these networks is not very clear, very well-defined, always changing quite like traffic ... The goal is to get networks to be able to sense themselves, their operation, and learn from their operation, and then to control how they operate but automatically without human intervention. That is our research."
Prabhakar and Fay share what's essential to making these networks work: synchronization.
"So, obviously, there's some technical challenges to achieve this, and maybe talk a little about high precision clock synchronization because obviously that is a foundational technology that would be required and is required for these types of technologies ultimately to be successful," said Fay.
As Prabhakar noted, data centers were "developed with the goal of taking big jobs but getting them executed on many small machines, as opposed to a single big supercomputer," meaning that as soon as something is distributed, all those computers have to act as one, like it's a single entity.
And now? These networks can coordinate down to the tens of nanoseconds, which is essential for ensuring fairness for financial systems, making sure that trades are executed in the order to which they're submitted.
Listen to this episode of Tomorrow's Capital to hear more about self-programming technology, as well as cloud computing, distributed ledgers, the Internet of Things, and much more.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
