ADLINK: Giving Back to Open Source — with Angelo Corsaro

Gael and Thabang are joined by Angelo Corsaro, chief technology officer at ADLINK Technology, to discuss the open source edge computing landscape, and how open source participation is a two-way street.

Transcript

Copy Transcript Copied

Thabang Mashologu:
Hello, and welcome to the entrepreneurial open source podcast. A show where we discuss the power and practical value of open source in business and entrepreneurship with leaders and innovators from around the world. I'm Thabang Mashologu,

Gaël Blondelle:
And I'm Gaël Blondelle.

Thabang Mashologu:
We have got a really great show for you today. Our guest is Angelo Corsaro. Angelo is the CTO at ADLINK where he leads the advanced technology office and is responsible for corporate technology strategy and innovation. Welcome to the show Angelo.

Angelo Corsaro:
Oh, thank you very much. Very happy to be on the show and I'm looking forward to this podcast.

Thabang Mashologu:
Awesome. Can you start by telling us a bit more about your role at ADLINK?

Angelo Corsaro:
Sure. So I've been serving as a chief technology officer in ADLINK since roughly 2016. When I think ADLINK as a company, decided to play a major role in software or in a way to be more present in software. Because as you know, we are a big outdoor company. Manufacturing anything that goes from small board to serve that you find in your cars air, traffic control, maybe telco and so forth. But back then with the race of edge computing and industriating of things, it was the realization that ADLINK was already serving many of the verticals and the market that would have been more impacted by the transformation that is being brought by edge computing. And the other realization that we did communication and software that helps in general. Sharing information was going to be one of the major elements in catalyzing these kinds of innovation and the application scenarios that are often envisioned for edge computing in industriating of things.

Angelo Corsaro:
This is why back then the technology office was created. I was asked to lead it and serve as the CTO of the company. And as I think now it's quite visible on the market overall, we started really driving and innovating in few areas in specifically everything that has to do with how you deal, we did the distribution over a very large scale in a completely decentralized manner, which you say a central problem in edge computing. And the other is how do you make very efficient communication also locally, let's say, within a car, within a medical device. And then the other is how do you provision and manage a system that span from the cloud down to the microcontroller? So this is our focus I think the reason why we decided to really invest and innovate in this area and I think that technology that we are incubating as they are they're open because we are a strong believer of the power of innovation that comes with open stores and strong believer of the fact that in a way, if you want to accelerate a domain, it's much easier to do that by establishing an open platform.

Thabang Mashologu:
Yep. Fully agreed. So how did you first get involved in source personally?

Angelo Corsaro:
Myself, I have a very long history in open source. Well, I still was a teenager and I used to have... internet meant mostly dial up modems, if you recall those.

Thabang Mashologu:
Yeah.

Gaël Blondelle:
I do.

Angelo Corsaro:
And as a young teenager, what I found terrific about open source. And back then, I still recall looking at the Linux Kernel or with my brother building our own local network with coax cable. Back then you had to recompile the Linux Kernel, if you wanted to have a support for an internet card and set up jumpers on the card. So as a team for us, what was cool is that we could see code and learn from code that talented people at Britain. So I recall reading code from Linus Torvald. I recall reading code, for those of you that recall the gang of four, by Erich Gamma. So I don't know how many people recall that, but the very famous book, the gang of four or design patterns.

Angelo Corsaro:
Most of the framework that they designed, it was actually open source. And so when I was a young teenager, I think for me open source beside the theoretical courses that I was doing in my studies was just a way to accelerating learning. As I think without open source I could have not done. The amount of trick I learned by looking at code of other people is uncountable. And I think that open source still serves that role which is fantastic.

Angelo Corsaro:
But then as I grew and eventually I started working and eventually I started becoming someone that not only loves technology and loves to innovate, but also needs to manage innovation. Then I started to appreciate the role that open source has in facilitating what I call, and I think it's actually commonly called, open collaborative innovation. I think that's very powerful and there are several other aspects that we can debate that make open source very powerful as a mean to grow our understanding of how a computer works and how good software can be written as well as a need to innovate and manage innovations as a company.

Thabang Mashologu:
So we've established that you have a long personal history in open source. Now coming into your role at ADLINK a hardware company, did you have to convince people of the merits of open source collaboration and innovation? Were you a change agent in that regard?

Angelo Corsaro:
So I think at the top level. So if you look at our CEO, he didn't need to be convinced at all. He has a very clear vision and he understands, I think, the power and the benefit of open source and in a way that might sound cheesy, but what I like to think, and I think it's a metaphor for living that I take in my life also, if you want good things for you to happen, you have to start doing good things for the others. And that's how we see open source. Okay?

Thabang Mashologu:
Yeah.

Angelo Corsaro:
So it's a way of contributing very sincerely and in a very generous manner. What we think now is the solution for a given problem in a specific domain, which in this case is a relatively large domain. Whether it be called edge computing for company, but whatever you prefer. But I think the investment that we have decided to do over the past few years was very generous because we've put the most talented team of engineer that we have in our company to the service of the community. And obviously we use those techniques ourselves, but we are really trying to do that to accelerate what we think is the right paradigm for computing going forward and make sure that we can have a rock solid platform supporting that paradigm.

Thabang Mashologu:
Right. Makes a lot of sense. So for our listeners who may not be familiar, what open source projects is ADLINK involved in.

Angelo Corsaro:
So our three key project today are Cyclone DDS, Zenoh and fogOS. And so maybe I should spend a few sentences giving some context.

Thabang Mashologu:
Mmm-hmm.

Angelo Corsaro:
So cyclone DDS, for those of you that have heard the term DDS is a DDS implementation. What is DDS? DDS stands for data distribution service. It's a standard-based data sharing technology, to which I've contributed a lot in the past because it was one of the guy that started the standardization group. And which is used heavily in aerospace and defense, in air traffic control management and in robotics. So some of you might know that DDS is used as the default transport mechanism in the robot operating system too. So why did we start Cyclone? We were already a DDS vendor. But the world has changed. The technological landscape has changed. Our graphs changed and the use case have somewhat changed.

Angelo Corsaro:
So we decided to incubate Cyclone to make sure that we could capitalize for the experience that we had in 20 years as a commercial vendor of DDS to try to say, okay what have we learned in these 20 years? What has changed? And if we were to start today, what would be the perfect DDS implementation? And this is what we are trying to... The mantra that we have formed for Cyclone. And this is paying back because it was selected as a default in ROS 2. And it's gaining popularity quite quickly. So the other project that we have in eclipse is Zenoh. So the problem we tried to solve with Zenoh is merging state-of-the-art protocols technique for efficient data dissemination and state-of-the-art technologies for geo distributed query to make sure that with a unified set of obstruction. And we have a handful of obstruction.

Angelo Corsaro:
Because in the end we found out that we need a publisher, subscriber and the concept of a queriable besides that of resource. And with these, we are able to model pub sub queries over joe distributed storage's and distributed computation. So all of a sudden we're able to make it possible for you to store data wherever you want using your favorite database. Because by the way, we do not want to create new databases. There are so many great database out there. But what we want to make it possible is for you to decide where conceptually and because of the specificity of your system makes sense to store the data. Perhaps you want to keep it closer to the computation right? And then in a complete location transparent manner, be able to clone the network when you should query, get the data from these judicious storages and get it back to you.

Angelo Corsaro:
So, as far as I know, Zenoh is the only protocol that does that. And I think it's a project that is growing very quickly. We are quite happy of where we are at. And as a super young project, because Zenoh is technically one year old, one year and one month in three days. So we can still count. And we have actually found that in challenging use cases like the tele operation of the Indie Autonomous Challenge car. And we are far more example of which we can't necessarily talk yet. But it's promising, it solves a problem that isn't necessarily solved yet by the technologies. And it's embracing because one of the things that we do for instance with robotics is that we have bridges from DDS to Zenoh that allow you to very easily, all of a sudden control robots from anywhere on the network. Or if you have a swarm of robot, use Zenoh to do mesh communication and so on and so forth.

Angelo Corsaro:
And then the third project that we have is fogOS. And you can think of fogOS as a CougarNet is, but with a difference that instead of provisioning containers, it has been designed so that it can provision just about anything. And why so because in our world, sometimes containers are too heavy weight and we need to provision maybe a uniquer or maybe a binary or maybe a library. So we have this concept of a fog deployment unit, where a fog deployment unit can be just about anything, of which we have implemented a plugin that allow us to manage the fog deployment unit. And then your application happens to be a graph or fog deployment unit that we can deploy spanning from a data center down to a micro controller.

Gaël Blondelle:
Maybe just to continue on this idea of edge and fog, et cetera. We meet regularly, Angelo, you and I, you and me in some meetings in Europe about edge computing and IOT about far edge, et cetera. So could you share with us your vision about how open source can help in this domain and can contribute to such successful deployments of edge in IOT and fog whatever you want to call it.

Angelo Corsaro:
Yeah, absolutely. So, and again, this ties back, I think, in a way to one of the questions that was asked before, which is why is ADLINK running this many open source projects in each company? So, as I said, from our perspective, there are 1 million good reason why going forward? We need to have more decentralized architecture. There are privacy concerns. There are energetical implication. There are latency. We could stay an hour here to debate why that makes sense, but we have debated that in the past so I'll give it fore granted. So now, if we all agree that edge computing is good, then the point is okay, but to do that, we need an infrastructure. If you look at what happened with cloud computing, one of the aspect that made cloud computing flourish was the availability of open source infrastructure that you could leverage.

Angelo Corsaro:
For edge computing we didn't have those. And so one of the reasons why as a company, we decided to really try to collaborate without an industrial partner to establish a platform for fog edge computing was precisely because we think that if we want to innovate, if we want to catalyze a domain, you need an open source platform. That's the first role of open source, you have that foundation infrastructure where now we can really start discussing about building applications that embrace these, at least parallelly. But that's not enough because then there is another aspect which is standards. And again, I've spent almost 20 years working on standardization up to when after 20 years, I maybe I got tired because in a way it's repeating, and I justify back to myself. I saying it's all the standards. But anyway, I stopped working on standards because from my perspective, standards are not easy to reconcile with the dynamic of the technological landscape we live into today.

Angelo Corsaro:
So my experience was that 20 years ago, we would do a standard. And that standard would became relevant for a few years, maybe even 10 years. Today, if you do a standard in our domain, it's obsolete by the time it's out, which is unacceptable. Because you start with a problem. It takes you two years and in two years, so many things have changed that it's obsolete. And so from my perspective, I've no interest in dedicating some of my time to that activity anymore. And the way in which we need to agree on how to do things is by open innovation and open source. So, as you know, in Europe, lots of discussion on what is the standard that we should adopt. Well, I think we should agree first on the technology, making sure that technology is open and we have an open innovation, let's say, ecosystem in which we can collectively innovate and embrace.

Angelo Corsaro:
So I think that's far more powerful and it's a better way also of directing European funding because it introduces less fragmentation. Because if we say this is the standard, and you have to do point implementation of that, you're fragmenting investment. If we say this is the open source project where we all invest, we all innovate. And this becomes our common circle, our common ground. It's far more powerful. And I think that's where we need to go as Europe. And this is true in general and is used in industries. Because that's what industry do for real. You collaborate on what is commodities and what is common, what is infrastructural, and then you narrate on top and that's where you differentiate. And I think we need to learn to do that also in innovations in Europe.

Gaël Blondelle:
Very good point. I love it. Maybe just to come back a bit in what you said previously, at some point you said we use open source to manage innovation. And also, like you just said to... for... Because that enables open collaborative innovation. And then if I link that to your story about DDS and Zenoh and then the adoption in ROS 2 the question maybe is, was it easy to get included in ROS 2? Because sometimes people think that I will publish my software in open source and then magic happens. So was it magically happening or was it the result of some strong investment from your side to work on integration with ROS 2?

Angelo Corsaro:
Well I think the answer is there is no magic. And there are no secrets it's... For any of you that have seen Kung Fu Panda. I love that cartoon. Because there was this thing that if you were elected, you could get access to the, the secret writing. And then the secret writing, the secret is that there is no secret. So if there is no secret, then you should ask what is the secret of success? But it's hard work. Dedication.

Thabang Mashologu:
Yep.

Angelo Corsaro:
That's what it takes. So do things that you love, do them extremely well, be sincere to your user and be supportive. That's what we do. If there is a secret, this is our secret. We love what we do. We love to hear and to solve problems from our user. And we take our users in general, we take responsibility for them. And anybody who has interacted with our forums, they will notice that we tend to be... I think we are very welcoming. Anybody posts a question, if it's in our time zone in half an hour after, at most you have an answer.

Angelo Corsaro:
And again, any issue we take it seriously and we try to solve in the shortest amount of time. I think that's, that's the only secret if there is a secret.

Gaël Blondelle:
Good. I love it.

Angelo Corsaro:
But this way of saying is like often it's 1% inspiration and 99% perspiration.

Thabang Mashologu:
And what would you say is the impact of DDS being made the default middleware for ROS 2 both, to the ecosystem and to your business in particular?

Angelo Corsaro:
So there are really, I think, two aspect in your question, because there is the adoption of ROS 2 of the standard DDS as the communication mean, which I think that gave DDS unprecedented visibility because DDS was really a niche of things before. And it's used in very complicated, complex and fascinating systems. Just few people know that DDS is used the rocket launch system of the Kennedy Space Center, as an example. It's used in air traffic control management, but few people know. ROS gave DDS huge visibility. So that made it to be a standard. Then for what concerns our implementation, that gave Cyclone visibility. It gave Cyclone an humongous user base. Because the level of testing that now we get because of ROS is humongous.

Angelo Corsaro:
And as in Spider-Man with great power comes great responsibility because now we have to be super careful at all changes we make sure that we maintain always backward compatibility, withdraws and so on and so forth. So that's an additional constrain to make sure that again, we are very good citizen in that domain. There're new requirements that are somewhat specific to robotics, which is very nice and we are happy to embrace, but again, the way in which it gave visibility operates at two levels because it gave visibility to the DDS as a technology overall, and then some visibility additionally to Cyclone indeed.

Gaël Blondelle:
So you answer each question with a quote from a movie.

Angelo Corsaro:
Yes indeed.

Thabang Mashologu:
I think it's awesome. Keep it up. If we can switch gears slightly, we've talked about the technology strategy and your contributions and participation in open source. Can you give us a sense of how do you monetize your involvement in open source? How do you pay the bills?

Angelo Corsaro:
So I think that, thank you for the question, because that's very important because some people think open sources for free. So just use it and abuse it and enjoy. Even water, some people think it's for free, but it isn't. Because something that is for free doesn't mean that you shouldn't pay attention or you shouldn't contribute back one way or another. Right?

Thabang Mashologu:
Right.

Angelo Corsaro:
And so I think that there is sometimes, which is a pity. Some people have, we call them free rider, an abusive approach open source, which is these people are like crazy guy that only are computer geeks. They spend their life in front of the computer in any case, that's how they have fun. So it's fine if we just use it and don't give anything back. So I think the responsibility of a good open source citizen is contribute back. And contributing back with your possibility, endorsement, that's fundamental. So if you use something, you should endorse it because that's good for the project. It's funding. There are multiple way of funding. And I think it's very important. We do that, right. We fund developers from other project. So I think that's very important and it's a matter of being a responsible citizen. Because as we know how much it costs, you're employing very talented people and making everything available. If everyone is responsible. That's something that will become far more sustainable.

Angelo Corsaro:
Then that is the aspect of responsibility. And I think that has to come with education because nothing is for free. And if there weren't company investing. Those things wouldn't exist. The company that invest as you properly mentioned, we need to find a way of monetizing. So in our case, we are somehow privileged because we sell hardware. And let's say that sometimes software helps us selling little bit more hardware. Which is a good way of indirectly monetizing our investments. And I can give you some example. Some people adopt our software, then they realize, but they also sell hardware and we need this, this and this. So they have a great experience with us like a company is natural. I like to make business with people I trust. I've developed a good relationship.

Angelo Corsaro:
So that's very natural. And that's an indirect way of monetizing our investment in open source. And then obviously some people want support contract. And again, it's very hard to build a company just over support contract. Because again we have had this debate in my past, in other company quite often. I recall some people complaining the way in which we support the community is too good. We have to be less good. Otherwise, people won't want to pay for support. And for me, that's unacceptable because it goes against the principle of open source. At the same time, if we want to maintain those principles, as I said before, we need people that understand the need to be good citizens. And again, where the support coming into the picture is for a project that had very long lived deployment.

Angelo Corsaro:
Okay. So that's where you can get somebody's session. And then if you look at modern days, the best or the easiest way of monetizing software is if you can provide some platform as a service solution around it. But that's not easy for all kinds of software, obviously. The part of monetization is complicated. It's not simple. There is not one size that fits all. I think it also depends a lot on the domain you work in because if your technology, and there are some domains in which it's easier to monetize than others. That's for sure. And that's usually where we focus for monetization. But I think overall as a community, I think we really need to educate open source user and make them a little bit more responsible with respect to this giving back.

Thabang Mashologu:
So Angelo, what would you say to somebody who is listening to this podcast, they've done their research, or they're doing their research and they're considering an open source based strategy? What would you wish somebody had told you 10, 15 years ago in this industry?

Angelo Corsaro:
That's a complicated question. So the, the first a warning that I would give is open source doesn't mean just open up your positer and get up because that's what some people think. So [inaudible 00:28:24] isn't actually open source. Then some other people says I have a repository and an OSI compatible license. So I do open source. Well, no, so for me, open source is really, or building open sources is meaningful if you build a community. So open source is about a community. And with a community comes a clear governance, it comes a responsibility. Because you have to talk, you have to listen, you have to support your community. And that is really the part that in one way is hard. It's time consuming, it's effort, but it's also very rewarding. So my advice would be, well, if you're just trying to open source your software and put it available just for everyone to use it, just think twice, because the real power of open source is building a community.

Angelo Corsaro:
Because that's where you get the give and take. And again, if you don't try to build a community, if you don't put the effort to build the community then it becomes like a monologue. Okay, my software is there. You like it, you use it but, for me, it's the fact that open source is just an open source license. That can be more or less, let's say restrictive, but it's not very interesting. The real power of open source is the community. Because it's the way of innovating together, it's the way of making sure that you have technology provider and user that come together to facilitate also user driven innovation. So I think that's what people should be looking for if they really want to open source.

Thabang Mashologu:
Yeah. Fully agreed. Now, what would you say has been your experience working with open source foundations, as far as cultivating that community, building that type of interaction and engagement?

Angelo Corsaro:
Our experience before deciding which foundation to join. We looked into several of those and in the end, as you know, a few years ago, we decided to ost and commit to eclipse. And so we had done our due diligence. And what I can say is that the due diligence was positive. And more importantly, our experience now, after several years of being in eclipse, a low season we ar very happy of the support, we are very happy of the way in which the community works and collaboration between different projects. So yes, I think we are happy of our choice and we are happy to contribute to eclipse. And we are happy to contribute in making eclipse in even bigger and more successful thing.

Thabang Mashologu:
And we are happy to have you and very grateful for the impact. Now, before we let you go, are there any exciting examples of usage, scenarios and implementations that you can tell us about?

Angelo Corsaro:
Yeah. So, as I said, we have right now, the interesting thing is that for one reason or another, and I don't know if it's necessarily related to the ROS 2 adoption. And the fact that I know by Cyclone, we are the default and the Zenoh as maybe some of you have seen, solved some robot verbal communication problem very, very well, but we've seen lots of adoptions in, in robotics. So we have some external adoption. We have some internal adoption, by the way, I'm not sure many knows, but ADLINK and FoxCon have started a few company in robotics.

Angelo Corsaro:
And one of these is called FA Robot. And we focus on AGB's. So those AGB appear on our entire communication stack. So they use the cyclone they use fogOS they use Zenoh for swarm communication. So it's good when you eat your own dog food. Internal adoption it's much harder than external adoption. So in terms of external adoption, as you know, there are some I can talk openly. You have the Indie Autonomous Challenge. There are a few others but I cannot talk openly, but I think are even more articulated and hopefully we'll be able to have some press on them quite soon.

Thabang Mashologu:
All right. Well, we look forward to that. I think that's a great place to leave it for today. Thank you so much, Angelo, for coming on the show, it was a lot of fun having you on the podcast. That's about it for this episode of the entrepreneurial open source podcast. You can find us online at entrepreneurialopensource.com and @OSS4Biz on Twitter. We'll catch you next time on Spotify, Apple Podcasts, or any other place you listen to podcasts. Thanks for listening.

Thabang Mashologu:
The entrepreneurial opensource podcast is sponsored by the eclipse foundation. The eclipse foundation provides a global community of software developers and organizations of all sizes across industries with a business friendly environment to collaborate on open source software innovation, visit eclipse.org to learn more.