Inside the Accelerator — Kings Distributed Systems
The Catalyst Cyber Accelerator is the first cybersecurity-focused commercial accelerator in Canada. Headquartered at the Rogers Cybersecure Catalyst at Toronto Metropolitan University, the Accelerator offers technical, strategic advice, mentorship and business resources to cybersecurity businesses that are ready to become national and international competitors.
In August, the Accelerator launched its third cohort with five innovative Canadian cybersecurity businesses. In this series of articles, we’ll be talking to the entrepreneurs behind each business about the Accelerator experience.
Kings Distributed Systems (KDS) has developed the Distributed Compute Protocol (DCP), a virtual computing platform to power data analytics and AI/ML applications in healthcare, smart manufacturing, cybersecurity, and advanced research computing. DCP allows institutions like hospitals and research centers to harness the computing power found in all of their idle computers and devices to run their computational workloads. It creates on-premises distributed computing clusters that require no incremental capital expenditure, are 10x cheaper than commercial cloud, and keep all data private and on-prem.
We spoke with CEO Dan Desjardins about the company’s rapid growth, its time at the Accelerator, and how their “Distributed Compute Protocol” technology works…
Kings Distributed Systems is built on a technology called the “Distributed Compute Protocol.” What is the technology?
My background is in physics, and I do all kinds of things with electromagnetic propulsion — how magnets interact with conductors and generate forces and fields, and all the equations that predict what those forces are. Those equations take a lot of computing power, and at the time I didn’t have enough computing power, or the skills to go and use a supercomputer. The Distributed Compute Protocol harnesses all of the idle computing power in computers and devices around us and makes it available — ease-of-use baked in — to do more computational workloads. I needed compute for physics, while others may need it for A.I., data science, blockchain, looking for tumors in MRI images, and so on.
What has been the evolution of the business?
In 2017 in a barbecue setting with a couple of friends, two colleagues said, “Hey, do you know stuff about compute?” I said, “Yes,” and they said, “We’d like to pick your brain.” Shortly thereafter I was #3 in this startup that was on a mission to build a distributed supercomputer. Fast forward four years and we’re now 32 employees. We’re building applications for smart hospitals and smart manufacturers and in growth mode.
How is Kings unique in the marketplace? Also, how exactly does the product work?
The Distributed Compute Protocol is one of many distributed computing platforms. There’s nothing novel about the idea of taking a computational job and spreading it out into many pieces to make it go faster. In fact, the very next thing that came after computing was distributed computing, way back in the ’60s and before. But what we’re doing that is completely different from, say, virtual machine and container deployments is that everything is built on web technology. If there’s one technology that runs on every single device you can imagine, it’s the web stack. Your iWatch, phone, tablet, even your fridge with the screen in the door, and definitely your laptops, computers and servers will run web technology.
We made a design choice a couple of years ago to build this distributed computing all-batteries-included framework on web tech, which means that it seamlessly and securely runs everywhere. It’s interoperable and compatible with everything, which is not true of many of the other technologies. I’ll give an example: virtual machines have high overhead and therefore need to run on beefy workstations or servers. Containers are great and lightweight, but they share the host machine’s underlying operating system, which means applications that target specific operating systems will only work in containers on hosts that have that operating system. DCP is both lightweight and will run on every single device out there. It’s combining the best of both worlds, and it comes with security and performance baked-in because over the last two decades, Google, Apple, Amazon and Mozilla put billions of dollars into making web technology secure and fast. We’re building on top of all that.
Who are your clients and how do they use the technology?
Clients are in the healthcare industry and the smart manufacturing industry. Both have the common theme of wanting to process a lot of data, use a lot of predictive analytics to optimize their operations, but both are hesitant about relinquishing a lot of their data to a foreign cloud provider. They’re also hesitant about paying that absurdly high cloud bill every month, especially when there’s a high compute involved with big data and large workloads.
What brought you to the Catalyst?
The very notion that data security and cybersecurity are absolutely paramount to developing a technology like the Distributed Computer Protocol. We’re talking about deploying software agents across all of the IOT devices and computers behind an institution’s firewall. It is absolutely critical that systems like the ones we’re designing are secure and compliant with all the existing frameworks, because our job is to protect people’s data and computed results, and so it’s paramount that our systems are leak-proof and secure. Through the Catalyst, we’re meeting expert over expert who are giving us all kinds of excellent insights to make sure that not only are we ready to scale, but we’re able to scale securely with customer data security in mind. It’s been invaluable to meet all of the experts that have been brought together under the Catalyst umbrella to give us excellent advice in how to focus cybersecurity and privacy aspects in the right places within our product and within the company.
Is there anything specific that you’ve learned that has already helped you?
Several things. There have been collaborative opportunities — for example, fusing certain of the cybersecurity sensing technologies on the market today from big companies with our compute network technology can yield dual-function compute networks that can analyze cyber threats in buildings themselves without leaving the customer’s firewall. Rather than exfiltrating network activity to a data lake in the cloud in the cloud, all of the analytics can happen within the context of the customer’s network infrastructure.
There are other specific insights with respect to providing layered access to different resources — data, code and compute — within our organizations to make sure there is no leakage of data across users even within the same institution. So there have been technical insights, policy insights, and collaboration opportunities with the Entrepreneurs-in-Residence, as well as the external third-party experts brought in by the Catalyst.
Shout-out to Mandy Bachus, who’s been excellent walking us through the marketing positioning materials. Shout-out to Erin McLean — excellent at giving us some very macro-strategic advice. Michael Ho has been formidable at helping us with his priority-mapping framework. RBC has been very helpful, and Palo Alto Networks has been generous with their advice and follow-ups. And big shout-out to the Catalyst team who put together an absolutely formidable program. This is by far the best Accelerator program that Kings has participated in.
What do the next five years look like for Kings?
Well, the next one year will be $12 million of revenue in the healthcare and automotive and manufacturing industries. Within the next five years will be a fully self-serve compute framework, built on interoperable web technology. We’re aiming to grow to 250 employees and $2 billion of revenue, and we will be the web standard for distributed compute.
