Why We Donated Project EVE to the Linux Foundation

by Zededa
May 14, 2019

Whether you believe in “Software Eating the World,” Digital Transformation, Industry 4.0 or an impending AI Revolution, you must agree that we are living in a world inhabited by “smart physical objects.” The behavior of our doorbells, cars and oil refineries is now dictated by the software that drives these devices. Of course, making generic hardware more valuable with software is nothing new. An Android phone is a great example of that. What started as a generic, fixed function piece of hardware in the ‘80s became a powerful mobile computing platform the moment we started putting software in the driver seat.

So, why all the recent excitement about hardware becoming “software-defined”? Are we just iterating on decades-old ideas, or are we truly talking about a significant shift in capabilities? I believe it is the latter. The advantage of “smart” devices—which can receive new and updated software automatically through network connections—is that their functionality can be adjusted, made more secure, and improved over time. Now, smart devices can increase in value rather than being depreciating assets. Being able to upgrade without swapping out hardware is a big advantage in the connected world of IoT.

I first came across this idea while reading Kai-Fu Lee’s AI Superpowers. He observes that prior industrial revolutions were significantly slowed down by the generic hardware R&D and adoption cycle. For these previous tectonic shifts in technology, “…physical products had to be invented, prototyped, built, sold and shipped to end users. Each time a marginal improvement was made to one of these pieces of hardware, it required that the earlier process be repeated, with the attendant cost and social friction that slowed down adoption of each new tweak. All of these frictions slowed down the development of new technologies and extended the time until a product was cost-effective for businesses to adopt.”

Contrast that description with what happens today when you buy a Tesla car. The physical device itself, of course, has to be there and it has to provide a certain amount of hardware “building blocks” (frame, seats, steering wheel, etc.). The software then orchestrates those building blocks into increasingly valuable features so that, five years and a dozen software updates later, you have a much more valuable, partially self-driving car. As you can see, your “self-driving car adoption cycle” is very different from what your great-grandparents’ “gas engine adoption cycle” used to be. You don’t have to wait many years for a big-bang arrival of a self-driving car—you can buy a “smart car” today and let it become a self-driving one over time.

The benefits of “smart physical products” to consumers and business alike are enormous, so how can we start building more of them as quickly as possible? The answer is two-fold. First of all, we have to make hardware itself flexible enough by providing easy-to-orchestrate hardware capabilities. On top of that, we have to make sure that software representation of those hardware elements is flexible enough to allow for a rapid software iteration cycle and safe experimentation.

To put it another way: we have to allow software to work with virtualized representations of the hardware elements.

If that sounds familiar, it should. Virtualized representations of compute, storage and networking hardware is exactly what we had to invent over the past two decades to arrive at public cloud and utility computing. It turns out, that was just the beginning. It is now time to virtualize all real-world hardware elements, not just computers inside of data centers.

To contrast this concept with virtualization of a classical server, the new term that’s becoming en vogue is “Edge Virtualization.” We believe that designers of “smart physical products” need to move to this abstraction ASAP if they want to build products that are more like the Tesla experience than a Ford Model T one.

Unfortunately, OT (Operational Technology) and IoT systems are often still stuck with antiquated software infrastructure and development techniques. These systems still behave as part of the traditional hardware adoption cycle, ignoring all the wisdom accumulated in the past few years that has allowed companies to innovate at faster speeds than ever before, thanks to cloud infrastructure and cloud-native development techniques. Right now, we’re still seeing systems that:

  • Have limited-to-zero remote management capabilities
  • Are often wide open to attacks, insecure and easy prey for hackers
  • Can only run one application
  • Are difficult to update or patch
  • Are based on proprietary hardware and software stacks

Virtualization, software-defined infrastructure and now containers have solved these issues for data centers and public clouds. We at ZEDEDA saw that OT and IoT system designers would benefit from similar technologies, but tailored to the kinds of devices and applications they are building.

To achieve this, we needed a common, ubiquitous piece of infrastructure for all the smart devices in the world. We need it to be open source and openly designed by the community of leading IT, OT, industrial and cloud companies.

We needed the Edge Virtualization Engine, known as EVE, to become part of LF Edge.

Project EVE is now open for business under the umbrella of the Linux Foundation’s LF Edge organization, and we invite everyone to help us build the future together – whether that future is called “Software Eating the World,” Digital Transformation, Industry 4.0 or an impending AI Revolution…

Get involved today at https://www.lfedge.org/projects/eve/.