Getting a tech innovation to work flawlessly in the lab is a massive achievement — but that’s just the beginning. These solutions are meant to be deployed in the real world, after all, and it’s much harder to control all the variables in new and unpredictable contexts. It’s particularly challenging for founders developing smart digital products that require both high compute power and low latency.
When Seth Hardy, co-founder and CTO of precision agriculture company Bug Mars, was first testing his smart digital platform, the conditions were far from ideal. The platform uses AI, computer vision, cameras and sensors for near real-time monitoring at insect farms. But at the test site — a farm near Kingston, Ont. — the Wi-Fi connection was so spotty, Hardy was unable to secure a connection more than a foot from the device. He spent long, frustrating days climbing up ladders and crawling around the insect barn, re-routing wire from the farmhouse out to where he needed it. “It was an adventure, and not the greatest one,” says Hardy. “If we’d been able to just ship them a modem and tell them to plug it in and we’re good to go, that would’ve been very different.”
The ability to stream large amounts of data in remote areas and analyze it quickly is essential to Bug Mars’ platform. Disease can spread quickly in insect barns, wiping out an entire colony within a short period of time. “You could leave for a day and your 100 million crickets could be all dead,” Hardy adds. “You need to have low latency — without any lag time. Otherwise, you’re going to be missing crucial information.”
A wide variety of innovative new solutions is dependent on having reliable access to high compute power and low latency. It’s what will help Calgary-based neurotechnology company Neuraura provide an immediate response to the health needs of its users. It’s also crucial to helping autonomous vehicles safely navigate routes, providing near real-time updates on weather events and enabling growers to monitor crops remotely. If it takes hours to process the information, the outcome, as Hardy notes, can be disastrous. But Wi-Fi, wired connections and 4G for data transmission don’t always cut it, and access to other services can be costly.
Ian McCarter, director of corporate innovation at MaRS, has seen the impact this can have on startups. “The testing phase is so important. But having access to technology to test and refine solutions is an expensive and often out-of-reach proposition for founders,” he says. “That can impose real limitations. The risk is that many of these great solutions won’t be able to commercialize fully.”
To give startups the opportunity to test their products on 5G and multi-access edge computing (MEC), MaRS partnered with Bell and Amazon Web Services. Through this special initiative, Bug Mars and Neuraura were given free access to AWS’s multi-access edge compute solution, Wavelength, which sits on Bell’s 5G network. And Bell and AWS provided troubleshooting advice, helping the ventures select the right technology and work through hardware design considerations.
“This project gave them the time — and the backup support — to test things out, which ultimately will help them build better products,” says McCarter.
Seth Hardy, co-founder and CTO of Bug Mars spent days rerouting cable for a pilot project. Getting access to a 5G network helped make testing his company’s product easier.
Struggling to find a steady signal at the Ontario farm, it became obvious that to capitalize on the potential of the company’s solution, Hardy needed to find ways to refine it — quickly. Over the past several years, the animal feed insect protein industry has experienced massive growth, thanks to insects’ high nutritional content and lower environmental impact. In 2021, the market for insects was more than U.S.$511 million and is expected to grow to about U.S.$10 billion by 2030.
But because modern insect farming is a relatively new field, “there are issues such as connectivity and networking that still need to be resolved,” Hardy says. “Some of these farms are in more remote locations. And smaller farms are often built in shipping containers, which act as metal boxes that block wireless signals.”
Through the partnership, Bug Mars was able to tap into a fast, secure network to test the performance of its platform. AWS’s Wavelength offloads part of a company’s computations from its cameras or sensors to AWS’s cloud, which is on the edge of the network. “The amount of computation that’s being done isn’t limited by the computational power of the computer that’s attached to the sensor,” explains McCarter. And because the data is being analyzed, processed and stored much closer to the source and without the use of public internet, MEC can improve the performance of the solution and eliminate the need for local hardware, reducing time and cost of customer acquisition while also maintaining security.
Those are particularly important factors in agriculture where the margins are notoriously thin. “Farmers are really concerned about having a competitive edge,” Hardy says, adding that the aim is to get insect protein to be cost competitive with soy. “It’s got a long way to go, but it’s coming down pretty rapidly.”
Plus, the technology helps provide an important layer of protection. “Having multiple points of global network connectivity access — whether it’s for redundancy or security — makes the operation of critical resources a lot safer,” says Hardy. “This technology helps build confidentiality, availability and integrity into our product and into the infrastructure in general.”
Neuraura’s health app enables users to track their symptoms with self-reported data.
Security is also a top priority for health tech venture Neuraura. It’s in the early stages of developing a prototype for a device that aims to help people manage symptoms associated with polycystic ovary syndrome, a leading cause of infertility in women. The company’s device will deliver localized central nervous system stimulation, building on research that shows that low-frequency electrostimulation reduces glucose, improves menstrual regularity and increases blood flow to the ovaries. Neuraura’s team is also developing an app that will allow the user to track their symptoms with self-reported data, and the company plans to integrate biosensors for hormonal and metabolic tracking as well as integrate data from other wearable devices.
To work effectively, the device needs to consistently collect and analyze data and personal information from the user, says Dharmang Desai, a data science specialist working with Neuraura on this project. The company was concerned that if data was stored in the cloud, it could be vulnerable to data leaks or hackers. But because AWS’s Wavelength processes the data and it’s not sent over the public internet, it is more secure. “Security is so important,” says Desai. “We are taking the users’ data directly, so it’s critical that proper data privacy is ensured.”
Through the program, Neuraura was able to set up its own data set on AWS; the company is now working to secure initial FDA clearance before it starts clinical trials.
This initiative brought the ventures one step closer to commercialization. “They were able to give us guidance on how to essentially extend our product’s infrastructure in the cloud to run on one of the AWS Wavelength zones, which can reduce the latency,” Hardy says, adding that while testing is early, “the numbers are pretty clear.” Bug Mars is now working on advanced disease detection through real-time video analysis.
For Neuraura, the partnership gave the team the opportunity to explore the up-and-coming innovations in the space, says Pierre Wijdenes, chairman and CPO of Neuraura. “It helps ensure we actually move the needle in a direction that is really forward thinking.”
Adopting breakthrough tech solutions means staying ahead of the competition. Become a MaRS Corporate Innovation member and gain exclusive access to our ecosystem.
Photos courtesy of Bug Mars and Neuraura