Local innovation hit the global stage at the Consumer Electronics Show in Las Vegas, Nevada this week, with the unveiling of Project Arrow, the first all-Canadian, zero-emission concept vehicle that has been fully designed and built here in Ontario, including integral computing systems from a Kingston company.
In a celebratory media statement, Ontario’s Ministry of Economic Development, Job Creation and Trade noted that 58 Canadian industry partners collaborated to design and build this revolutionary prototype vehicle. “Project Arrow is a true testament to Canadian ingenuity, hard work, and determination,” said Raed Kadri, Vice President, Strategic Initiatives and Head of the Ontario Vehicle Innovation Network (OVIN) at the Ontario Centre of Innovation (OCI). “We are cementing on the international stage that Ontario and Canada’s technology, innovation and manufacturing capabilities are world-class and second to none. OVIN and OCI are proud to play an important role in this vision.”
Among those 58 industry partners is Kingston tech company Distributive, whose open-network computing models are used in Project Arrow. Distributive’s CEO, Dr. Dan Desjardins, explained that the software they’ve developed has a wide range of uses. “The software that we build, called DCP, Distributive Compute Protocol, takes application workloads and spreads them into tiny pieces over all of the idle computers [in your network]. Say, you’re in a hospital, or a university, or government building, or a media office or a law firm or what have you, it uses your idle computers for computation, and then it returns the results. That should sound familiar that that’s exactly what the cloud is.”
Dr. Desjardins noted that Distributive’s model has significant advantages over the standard ‘cloud’ approach. “The cloud is [external] computers that you rent — you pay a lot of money for — and you run your applications and get your results back. So we’ve created software that essentially takes the cloud, breaks it up into pieces and does it over devices that are already in all of our buildings. Now, the advantages are that it’s 96% cheaper, because it’s using devices you already own. That’s important because cloud is super-expensive. Almost of equal importance: it keeps data in your building. Now you’re not sending data into the United States or to some data center 200 kilometers away or something. So for data privacy, especially in hospital contexts, where we are today, it’s very important.”
It may seem like quite a leap from office applications to futuristic zero-emission vehicles, but Dr. Desjardins noted that DCP is a great fit for Project Arrow’s high-tech computing systems. “These Arrow vehicles, they’re basically computers with wheels. You know, back in the day, you had these old cars that were, catalyzer, engine, combustion, drive train — a bunch of rubber and wire and metal. Not anymore. Now it’s more silicon than anything else. That’s a lot of compute. And so it has to have a lot of compute because we’re building so many compute-enabled applications, like automatic driving, steering, or navigation. So these cars are built on purpose to be true muscles in terms of compute, so that they can navigate.”
Personally, Dr. Desjardins was enthusiastic for Distributive to join the other Canadian companies working on the concept electric vehicle. “It’s Canada’s first EV. It’s an iconic project, and it’s called ‘Arrow’ which is great,” he said, reflecting on the history of the legendary Avro Arrow, another storied Canadian innovation. “I was a military pilot by background, so you know that the fact that was called ‘Arrow’, I mean, drew me in like a sucker right away… Now also, every car [being designed] on a go-forward basis is a moving computer. And so, Arrows are our starting point.”
Beyond the operation of individual vehicles, DCP’s technology has the potential to let the collective computing power of electric vehicles solve broader travel and traffic issues. Dr. Desjardins explored a hypothetical travel problem. “If you have 10,000 electric vehicles driving down the 401, and they all need to recharge for 20 minutes, somewhere between here in Toronto, what happens if they all stop at Oshawa? If there’s three people on Friday, that’s already an hour delay on your route. That’s crazy. So planning which car should stop where — this is called capacity management — that’s a big combinatorial problem that takes a lot of compute power. So imagine connecting all the vehicles to do that optimization with the vehicles themselves… Use the cars themselves to create clouds that do all the optimizations for not just capacity planning, or route planning, but also road condition, or predictive maintenance.”
The issues that the Arrow’s computers could work on while driving or idle do not need to be restricted to vehicle or travel-related problems. “They could be doing, like, scientific advancement in cancer research while they’re driving down the 401,” Dr. Desjardins noted. “It doesn’t matter what the problem is: computers compute. And here’s what we know, there’s an $800 billion need for compute, we know because that’s the cloud-compute market share. And it’s going to $1.5 trillion in 2029, fueled by AI and data science, everything else. There’s a lot of need for compute. There’s a lot of idle computers out there; Distributive is connecting it all together and serving it up 10 times cheaper, and making it easy for everyone to get involved.”
Beyond Project Arrow, Dr. Desjardins said that Distributive is busy working on other collaborative projects as well, including with a supercomputer community that engaged millions of computing enthusiasts around the world from the late ’90s onward. “So SETI @ Home, and later Asteroids @ Home, was one of the 20 projects that ran on this big global supercomputer. The underlying infrastructure [of the supercomputer] is called BOINC — Berkeley Open Infrastructure for Computing — and it ran many other different projects. The name of the computing community is called World Community Grid, and it was managed by IBM. But just this year, it moved. World Community Grid now falls under Krembil Research Institute out of Toronto, and we’re working with them. We’re actually on their advisory board.” According to Dr. Desjardins, Distributive’s software is going to be helping World Community Grid’s long-established community supercomputer become more versatile, and incorporate more locations onto their network more efficiently.
Cutting edge software is exciting, but Dr. Desjardins stressed that industry success “all comes back to people”, and he expressed gratitude for the support they’ve had along the way. “We’re very grateful for the people who’ve invested a lot of time, effort and faith into us. Because that’s what makes this work. We’re extremely grateful for the folks in Kingston who have given us the time to test these new cutting edge innovations because that’s how innovation happens.”
Dr. Desjardins laughed about Distributive being “an overnight five-year success”, noting that small companies take a lot of work and support behind the scenes to bring their great ideas to fruition. “Our one wish,” he noted, “is that more established companies and institutions would extend that same [support] to more Kingston startups. You know if everyone’s saying, ‘Oh no, no, come back to me when you have, like, 10 customers already with proven proof points.’ Guess what, if everyone says that, you never get anything off the ground. So I’m speaking on behalf of all Kingston startups – I’m connected to the ecosystem and that’s the most helpful thing – is when people offer time to try an innovation and give good critical feedback. That’s what gets companies off the ground. And I think together we can create a better innovative ecosystem in Kingston, in Ontario, in Canada.”