Proteantecs Brings Hardware & Software Together to Revolutionize Predictive Chip Maintenance

Proteantecs Brings Hardware & Software Together to Revolutionize Predictive Chip Maintenance

The last few years have seen remarkable advances in the cloud computing sector. Recent data suggests that over 90% of enterprises make use of cloud technology, representing industries from banking and education to healthcare and manufacturing, among many others. All of these advances have been accompanied by increasingly complex electronics, down to the microchips that make this computing possible. As the stresses on these chips increase, so does the possibility of a chip malfunctioning. And given how many people rely on these technologies in their daily lives—for everything from small conveniences to critical uses—that possibility of failure can mean greater risks than many might be prepared for.

But proTeantecs believes it’s found an answer to reliability concerns in high-stress, high-demand electronics. It relies on some of the latest developments in chip production and cloud-based technologies. But just as importantly, it involves a little extra listening.

Founded in 2017, proteanTecs’ team reflects a variety of fields: the staff includes software developers, machine learning experts, circuit designers, electronic design automation developers, and more. Noam Brousard, VP of Product, sees the company’s varied background as representative of its “best of both worlds” approach to chip design and maintenance. For Brousard and proteanTecs, crafting reliable, high performance chip technology means unifying the hardware production and software development processes.

In practice, the proteanTecs method means being able to understand how a chip is performing at all time by taking advantage of both sides of the equation. “We’re a hardware company with respect to the fact that we embed what we call ‘proteanTecs Agents’ in silicon,” Brousard says. These Agents are advanced monitors that measure various aspects of a chip’s health and performance. “But the other end of our solution is our cloud-based analytics platform, which actually receives this information and makes sense out of it, essentially providing actionable insights.” In effect, the proteanTecs approach focuses on “listening” to the chip as it operates, in order to obtain the most accurate information possible about its health. This process, in which chips self-report on their condition, is a major part of proteanTecs’ initiative called “Universal Chip Telemetry,” or UCT. Brousard hopes the idea will eventually become standardized throughout the circuit design and manufacturing sectors.

To illustrate the benefits of this approach, Brousard offers the example of a self-driving car—an autonomous vehicle operated entirely by AI. Brousard points out that the chips used in autonomous vehicles must be able to make decisions in dangerous and urgent situations, so there is no tolerance for chip failure or error. “Imagine,” he says, “that this artificial intelligence chip makes a decision that is life critical. For instance, it might decide if an airbag opens or the ABS [anti-lock braking system] starts working. The tolerance for failure is nil.” If there is a failure within the chip, responses typically involve trying to identify the point of failure during the chip’s manufacturing or usage by combing through the chip’s history. Though this might yield helpful information, it’s only possible after the fact.

With proteanTecs’ UCT technology, Brousard explains, the car’s chips continually communicate their status at each stage of their life cycle, allowing the company to monitor and produce diagnostics at every stage, including production, testing, and integration. “Using our unique ability to listen and provide visibility into the internal workings of the chip,” Brousard notes, “proteanTecs can alert users that a chip might fail three months in advance, rather than after the failure has already occurred.” This tech also allows proteanTecs to recognize faults during production, so potential issues can be addressed as early as possible. Brousard calls this “truly predictive maintenance.”

To make this kind of end-to-end monitoring possible, cloud computing is crucial, and AWS plays a key role in proteanTecs’ work. “Cloud is important in our life in many respects,” Brousard says. “The ability to run millions and millions of simulations at that development cycle is only possible if we have the ability to scale up and down with our computer resources.”

Brousard sees enormous need for this kind of work on the horizon as the demands placed on chips become even greater. “If you look at the sectors that need a solution like ours, it’s every technology that’s looking to scale massively,” he says. The companies seeking to use these highly advanced chips also require the highest quality possible, and they reflect an enormous range of industries. “We’re not trying to solve a corner case,” Brousard says. “We’re trying to start a revolution.” To guarantee the safety and stability of the technologies powering the future, proteanTecs starts by paying better attention to what that tech has to say.

from AWS Startups Blog

Sharing is caring!

Comments are closed.