So here we are at EclipseCon in Santa Clara and there's also the parallel OSGI thing running. And would you like to introduce yourself? Yes, please. I'm Chris Larsson. I'm the CEO of a company called MakeWave. Right. Based in Gothenburg, Sweden. Right. And we are the maintainers of this Knopflerfish OSGI distribution, which is one of the three open source, or four nowadays with concierge. Right. So what does that mean exactly? Are you in a run time for OSGI things? Yeah, we are a run time and we make a full implementation of these OSGI specs. Right. So we have it all. It also has a build system, of course. We actually have a plug-in for Eclipse, too, for developers. And we made it very early, even before the Eclipse guys were up to speed on that. In this sort of OSGI run time world, is it common to be able to move between each? Does the run time you choose make a huge difference in the way you have to go about your daily coding life? If you're a user of OSGI, it shouldn't. Then you've made a mistake. And many of the customers we have, they say they don't add extra features which are only available in your framework. Sort of proprietary features. Yeah, sort of proprietary. And I think that applies to everybody. And for instance, I know the Spring guys, or the interface, or Spring source they call nowadays. In the Spring framework, they make sure that they test always on Eclipse, or Equinox, on Felix, and on OperaFish. Right. It's sort of analogous to the old J2EE idea, that if you package things in standard Java Enterprise formats, you could deploy it to different application servers. Exactly. Right. And I think one key thing with OSGI, OSGI is not an open source project where the spec is the code, basically. Right. And then we have a well-written specification with different implementations that really tries out the specification as well.
2:00Which kind of extra adds an extra level and kind of makes OSGI a bit unique in the open source space. Right. And there's several commercial implementations of OSGI runtimes, right? Or are there only open source ones? There is. I think process still has their... Okay. Our implementation used to be proprietary, closed source, and then we open sourced it because we changed the business model. Sure, sure. And so do you guys just sell support around it and any subscriptions and things like that? Things like that. So we have support contracts, we have professional services, and we do training. So we're not unique in that kind. And so, I mean, it seems like, more than it seems like, it is the case that OSGI came from sort of like set-top boxes and embedded devices and mobile things. And nowadays, I mean, it's a huge part of EclipseCon that we get, it's on the server side, if you will. Exactly. And just more of a, more running on a computer rather than a mobile device. And I mean, how do you, I mean, you probably have some good ideas of how that sort of came about to be. I mean, how did it make, how did OSGI make a leap from running on mobile embedded things onto running on the server? Exactly how that happened, I'm not really sure, but the background is that we were, ten years ago, Ericsson started a project called eBox project. Uh-huh. And I wanted to put a residential gateway in each and everybody's home. And that residential gateway should connect all the devices. Yeah. And they should have a platform in which third parties could install software. And that, that software should be guaranteed to be, you should have integrity. It's like the computerized dream of a household. Exactly, right. A digital home or whatever it was called. Yeah. Yeah. It was really hyped around the, uh, $2,000. And, uh, for that purpose, they needed a software platform in this box. And, uh, we were assigned as a company, uh, to develop, uh, help Ericsson to develop this platform. Right. And that led to the work of the OSGI Alliance. So we've been very much involved in, in the first, uh, releases through R1, R2, et cetera, of the OSGI specs. And we had some, uh, some of the fundamentals designed there with integrity, bundles, et cetera.
4:01We were, we made them for a better device. And, as it turned out, those, those fundamental designs, they were missing in Java. We were clearly adding some benefits to the Java language in itself. And it, it has turned out that they are so useful also in the desktop environments and the server side environments. Yeah. And, but we didn't design, we didn't have a server side in mind when we designed this. It was, uh, I think we were lucky. We, we, we, we hit a, we hit a good spot in, in Java. Where there, obviously, was something missing where we added, uh, a very good value. What do you, what do you think those missing things are that, that OSGI brings? It's the, the dynamics of Java. That you can load and unload software, uh, on, on a running JVM. Right. You don't have to do this restart of the JVM to, to get a new model, module loaded or unloaded. Right. And also the concept of the service registry. I mean, we did a service registry before the term, uh, service oriented architecture were invented. So we, it, it, it was a kind of SOA, but within, within one, one, uh, VM. Right. I mean, that, that's, that's the, that's the analogy I use with people all the time. Is it, it, it's like an intra VM SOA that you're running. So, yeah. So what then happened was that, uh, well, the residential market didn't really take off. Right. As we all know. And the dot com burst and everything collapsed more or less. Uh, but the telematics industry picked it up. And they saw that in the car industry or in the, in, in the vehicles, it's a kind of similar environment. You have like one central unit and that, that is connected to numbers. Right. And, and, uh, so BMW launched, uh, project. I think they were the first to use OSGI and it's built, it's still under discussion. It's still used in several R and D products. We've done several R and D products in the telematics or automotive space. Uh, but they have very long lead times and they are a bit worried to step into new technology. Right. But I think eventually we will see OSGI like architectures in the vehicle. And, and, and sort of the, the user and customer base that you guys have. What do you find people are using Knoblerfish and OSGI for?
6:01Uh, lots of things. Just everything? Everything. That, that's the kind of interesting idea. Interesting thing for us. I think we, we come more out of the, um, uh, the embedded space. Yeah. Some kind of gateway inside a house. It can be some industrial equipment. It can be also telematics projects. But we're also getting now customers in the, in the server side space. And, uh, I mean, Equinox is a great platform, uh, but it's, it's kind of, has the IBM stamp on it. So we, we try to pick up the crowd that is not, that don't want to go in the IBM. Sure, sure. Like, it's like the phone industry you're talking about. There's a lot of political things. A lot of political things. So, uh, but it, but it's a tough competition. I mean, these days it's like, uh, the users say in the seventies, you don't get fired for, for buying from IBM. Right. Now it's like if you're a big IT company, you don't get fired for using Equinox. Nice. So it's, uh, it's, it's a tough competition. But on the other hand, I mean, they are, they are also marketing the technology for us. Uh-huh. So it's really two sides of this coin. It's a tough competition because they know it's IBM and all everything behind that. But at the same time, I mean, they are paving the way for OSGY and getting a lot of buzz about it. Yeah. Well, they're getting people excited about using it, right? Yeah, definitely. That's right. I think we're going to close the loop here as well because the, the residential market, I think that's about to resurrect actually. Yeah. Well, we are, we are kind of at the beginning of a cycle of, uh, people or want more devices. Exactly. Or, or not that they want them necessarily. I think people always want more devices. It's just that they, uh, they're continually disappointing. Like when you actually get a device, you know, they're kind of like a little creaky and like things don't exactly work. And so it's got to keep innovating. So it finally like just works and it's, you don't have to think about it too much. Sure. And, and hopefully we're at another uptick of a cycle where we're going to have another go at it, like just having devices. I've been thinking about what, what was the key factors why we didn't succeed like 10 years ago or eight, 10 years ago.
8:01And, uh, there was basically three factors. First of all, Java back then really sucked if you ran it on a device. Right. I mean, it was too big, too slow, very inefficient. Uh, you need lots of memory to run it, run on it. These days memory is not, uh, it's not costly with memory anymore. Yeah. That's interesting. I mean, I haven't thought of memory as like, it's, it's, everyone always talks about Moore's Law and CPUs and stuff like that, right? But when it comes to, to Java and even embedded things, it's kind of memory is like a pretty big blocker. And so, so, so it's cheap now. Yeah. I mean, it's been, I think that that has been the main blocker in telematics. Right. I mean, these guys, they, they don't count the megabytes, they still count in kilobytes. Right. Right. If they can, if they can send, save 10 cents by going down. Yeah. One lower chipset or whatever, they will do that. Right. Right. Then it was the, um, uh, the penetration of broadband. Right. 10 years ago, very few people had broadband or even fiber. I mean, we, we used dial up connections in the first projects we did with rarex. Right. So that has really changed the things. Uh, you have also the, the price of, um, of the gateway itself. Back then they were quite costly. I heard from a Taiwanese manufacturers of gateways that the, the, the, the bill of material now is one sixth of the price. Really? In the year 2000. Wow. That is cheap. That is cheap. So, prices have really gone down. And I think that the price level was also too high initially. Yeah. For these kind of systems. Yeah. And also, as you, there's a fourth thing, a fourth thing. There's the, um, uh, the devices were, were trying to connect. I mean, they were wired back then. Yeah. Yeah. You didn't have wireless networks. Well, you had some around, but not to the degree we have now. Yeah. Low cost, low on, on, on, on energy consumption. It's easy. Anybody can just get your kit. You have the gateway. You have a number of devices. Just hook them up. You just place them and they will work seamlessly out of the box. I mean, that's the, we can finally have a refrigerator.
10:01Tell us when we need to buy milk. Exactly. This dream will finally come true. Exactly. I think that that market is also growing faster in Asia. Okay. We're actually doing quite a bit of business in Asia. And they look at this at a different angle. And I mean, they are very keen on devices and they love new gizmos, basically. Right, right. Well, thanks. Was there anything else you wanted to mention or talk about? Not really. All right. Well, great. Well, I appreciate you taking the time. You're welcome.