EdgeX Foundry moves forward in its push to harmonize IoT edge computing

The EdgeX Foundry is taking another step forward in its mission to standardize IoT edge computing. The EdgeX Foundry, an initiative of the nonprofit Linux Foundation, is readying the launch of its first major code release, known as “Barcelona.” More than 60 EdgeX members, including the likes of Samsung and Dell EMC, worked on the code. The next release, dubbed “California,” is due in June 2018.

In addition, EdgeX Foundry and the Industrial Internet Consortium (IIC) have signed a liaison agreement that will unite the two organizations in identifying best practices for IIoT applications. The two organizations will also collaborate on experimental projects and work to improve the interoperability of IIoT technologies.

We want to help address this scenario. EdgeX wants to do for IoT on the industrial side what Apache did for websites.

To learn more about the organization, the Internet of Things Institute interviewed Philip DesAutels, Ph.D., executive director at the EdgeX Foundry and senior director of IoT at the Linux Foundation.

What is your guiding principle for the EdgeX Foundry?

Philip DesAutels: Let me put you in a way-back machine to the early days of the web. Back then, if you knew HTML and could do CGI scripting, you could write your own checks. But it took a lot of effort to make things work. After that, the industry all got commoditized very quickly — in a good way. Once we had Apache, it became very easy to set up a website. Windows Networking came along and made networking easy, too. Suddenly everybody had a website and e-commerce was a thing.

[Hear DesAutels give a keynote at Enterprise IoT World, an event that highlights the intersection of IoT and industry and showcases how IoT transforms business across manufacturing, supply chain and operations. Get your tickets and free passes now.]

When I think of the IIoT companies that are getting it right now, they are turning over crazy amounts of money like the people did in the early days of the web. They are like the people who first figured out how to do e-commerce or how to put a catalog up online before anyone else.

I hear that a lot of IoT projects are stalling at the proof of concept phase. What’s your take on that?

In the IoT space, we are at the roll-your-own stage. I can see why a lot of IoT projects are stalling at inception. Let’s say you are prototyping and you are working in Arduino and have maybe moved to some other type of code. Everything you built is fragile. Then, you go to the next stage, and you realize your project is trash. You have to rewrite your code. You realize that you only had 10% of the stack done when you did that first prototype. You got it to connect and send data, but you conveniently chose to ignore security, configuration management and device management and all of that. Then you say: “I have to build this stack on the device and a different stack for IoT edge computing and then hook it all up together and connect it to the cloud.” You might then think: “To make this work, I need to put 40 people on this project. But our entire IT team has 40 employees now. How are going to hire that many new people? ”

We want to help address this scenario. EdgeX wants to do for IoT on the industrial side what Apache did for websites. We want to make it really easy and create an ecosystem to support this initiative like with Apache. In the early days of Apache, all of the sudden everybody had a plug-in for it. If you wanted an image server, a catalog or a credit card processor, you could get them.

Without that kind of catalyst, every IIoT project is going to be bespoke. Or you have people just redoing the same stuff over and over again without knowing that someone else already figured that problem out before. On the other hand, with that catalyst, IoT becomes another bit of IT and OT.

We now have 60 plus members in EdgeX Foundry who are behind this.

Part of the challenge seems to be the sheer amount of variability in the IIoT industry, right?

DesAutels: I am seeing more models for IIoT. IIC is doing brilliant work there. They are capturing the models, and they are bringing them forward, so it is straightforward for industry to say: “OK, this is a good security model. Transport Layer Security is already vetted. We don’t need proprietary security because we have seen security that holds up to web scale.”

What are you seeing from IIoT technology vendors?

DesAutels: I love talking to the smaller companies that have a great service, whether it is a complex event processor or a security tool that does end-to-end security. And they say: “Our customers are buying our tool, but then we have to spend nine months with them working on systems integration because they need to integrate our thing with these seven other things. And two of those things change along the way because they didn’t meet that company’s needs. We end up selling mainly consulting and just a little software. But what we really want to do is sell software and sell no consulting.”

We need to get to the point where those types of companies can do what they dream, and not focus on holding things together in the middle. We want to be the glue that holds the middle together.

What do you make of the current protocol and standard landscape?

DesAutels: With IIoT, it is usually not like you have a new Wi-Fi sensor that you go stick someplace, and it just works. I have walked into factories, and some of their automation control is 50 years old. I saw a machine that uses analog computing and others that have serial lines coming off of them. The wires have obviously been there for decades. That is not a bad thing, but you do need interop.

Things are improving. Automotive is very quickly settling into standard models. The facilities management world has settled on Project Haystack as a common model.

When verticals get together and decide they are not going to fight and not going to make it about having a competitive advantage, the protocol soup disappears very quickly.

How is it working at the Linux Foundation?

DesAutels: I work with the guy who invented Linux and hundreds of companies that are all trying to create free stuff for other people. That’s magic.

It is one of the few places where you feel like you are working for the good guys.

How is it working with the inventor of Linux, Linus Torvalds?

DesAutels: He is part of the team, so if you have a question, you can call him up and ask him.

I talked to him a few years ago at the Embedded Linux / OpenIoT conference. Not to sound like a broken record here, but he looked at me and said: You have to learn from what we already did.

He pointed out that there are more web developers in Silicon Valley than firmware developers in the world. He asked: How are you going to make IoT work if you depend on firmware to create bespoke software for everything? That was such a punch in the gut.

What can you tell me about your farm?

DesAutels: We have close to a 40-acre farm. It is a full working farm. We have 250 apple trees and a bunch of other things. We have a bunch of co-farming with full-time farmers. It has been phenomenal.

But on the other hand, this ties back to open source: Last weekend, I needed to put up a bunch of posts. We are using the Espalier method — the trellis method — for apple trees. I was putting up 14-foot posts, but I didn’t want to bother my neighbors. But a couple of neighbors saw me and called some of their friends, and pretty soon, there were 12 people at my farm with three different tractors and a bunch of shovels. The posts went up in three hours. In the end, everybody just smiled, and they all just went away. They didn’t ask for anything. They just did it because somebody needed help. That kind of community in a farm community is so genuine, but I hadn’t thought about how close that is to open source and what I do in my day job.