The automotive industry was an “early adopter” of Edge computing which has only continued to grow and accelerate, according to Michael Maxey, VP of Business Development at ZEDEDA, who spoke to IoT Insider about how he saw the technology transforming this industry.
Early adoption of Edge computing
“The early adopters were the big industrials … these were the industries that were early to figure out, how do I save money? Or how do I do things faster by moving compute from a central location to where the data is created,” explained Maxey.
However, although industries like oil and gas and automotive were quick to recognise the value of Edge computing, they are no longer alone in deploying this technology, which has “really picked up in the last three or four years,” said Maxey. “IoT has been around for 20 plus years, so it’s easy to look back even further and say this industry has been ongoing for a long time.
“The difference in Edge computing [is that] you’re starting to see a variety of workloads and more specialty, whereas IoT tends to be on an individual device with an individual use case … Edge computing tends to be application based.”
What’s driven this shift? Maxey noted “security and dollars” as an ultimate factor in decision-making, but there is also one notable trend behind this transition to Edge.
“There are more and more things coming online,” he said. “If you think about your house, maybe your Wi-Fi would connect to your laptop … now it connects to your refrigerator and light bulbs and everything else. The same is true in the enterprise world. More and more things are coming online, and the goal of enterprises is to extract as much value out of these things.
“You could push it [data] to the Cloud and process it in the Cloud, but you’re going to have to pay egress fees and build very robust networks. And it’s not high-value data that you should be pushing to the Cloud … so instead of pushing the data to the Cloud, they’re pushing compute to the data.”
The added bonus of processing data on device edge is that it satisfies security concerns in offering greater protection to data as it doesn’t leave company premises, at a time when awareness around data security and privacy is significantly higher, owing in part to regulation.
Edge computing in automotive
What the deployment of Edge computing looks like in automotive focuses on, for example, improving the in-vehicle experience for end users.
“A simple one would be voice recognition,” said Maxey. “A lot of automobiles are bringing that voice-to-voice action. You can turn on your air conditioner or turn off the radio using your voice. That’s a simple, large language model (LLM) running in the vehicle that’s designed to understand [your] voice and take action.”
Another major use case is in servicing the vehicle: “How do you update the firmware on tens of millions of cars in an efficient way? With electric cars coming online, the number of software updates is increasing. So we’ve seen a lot of AI and data management not only in the car but through the whole ecosystem, all the way back to the factory.”
Thanks to more widespread adoption of Edge computing and the development of better tools and more available models, Maxey explained, the cost of building and deploying solutions has come down in price. “The entry point for an enterprise to use AI has come down significantly,” he added. “It no longer needs a billion-dollar use case. You can roll out AI on a million-dollar use case.”
Key takeaways
“The key recommendation I would make for people who are starting to think about this space is to get your data in order,” said Maxey. “AI lives on data and often most people aren’t collecting it. If they are collecting it, they’re probably not tagging and structuring it in [the right] way … all of that needs to happen before you can put an AI model in place.”
A wider problem of the technology industry can be “to gloss over the hard parts and focus on the shiny outcomes”.
“Number one, make sure you have a strong business case, because without a business case, you’re tinkering with no point. It needs to be a business case that will drive enough value that the executive suite will sign on to these types of projects,” said Maxey. “With that in place, spend time looking at options, there are a lot of companies out there … you’ll likely find something that was almost purpose built for your use case.”
In summary: clean up your data, make sure you have a suitable business case, and spend time finding the right solution for you.
There’s plenty of other editorial on our sister site, Electronic Specifier! Or you can always join in the conversation by commenting below or visiting our LinkedIn page.
