By Keith Basil, General Manager of the Edge Business Unit at SUSE
The far edge is opening up limitless opportunities for organizations from industrial factories to retail stores to collect massive amounts of data via Industrial IoT (IIoT) devices, and then leverage that data to improve their products and experiences. In fact, Juniper Research estimates the total number of IoT connections will reach 83 billion by the end of this year.
Consider a nationwide fast-food chain as an example. In any one of their locations, there are tiny Kubernetes cloud clusters that service and provide internet access to that location. On top of that, there are multiple IIoT devices throughout the store that serve various functions from measuring the temperature in the fryer, to counting the number of cars in the drive-thru, to analyzing the tone of customers’ voices as they order food. This adds up to anywhere from 50-100 sensors at each location. While these devices are too small to run full-blown clusters or Linux distributions, they do have software on them, meaning there are massive amounts of data being processed at every far-edge location.
The expansive growth of the far and tiny edge means organizations have gone from managing single-digit edge deployments to thousands. Teams working at the edge have had to completely overhaul the way they operate as a result of this significant change. Additional pressure is on developers in particular, as they are responsible for building and managing the software protocols used to connect edge deployments and devices.
Protocol Overload
As the far edge continues to expand and diversify, engineers are becoming increasingly burdened by the multitude of IIoT devices and protocols they need to work with.
Every sensor, from a pressure gauge in a factory to a thermometer in a fast-food kitchen, has its specific protocol. Luckily, a few industries have begun to standardize on one protocol. For example, most food and beverage providers use MQTT, while most smart buildings use BACNet. In most cases, though, countless new protocols continue to be developed, leaving developers to deal with data silos and increasing intricacy.
The explosion of AI in the last year has added an additional layer of complexity and urgency to this problem. Far and tiny edge devices are collecting much of the data that is powering generative AI, but the full potential of generative AI cannot be realized without connectivity to edge devices and the accurate, real-time data they’re collecting. Protocol complexity is adding another roadblock for organizations looking to leverage the power of generative AI.
Businesses and their engineering teams simply can’t keep up with the speed of change happening at the far edge. The immense scale of this problem is already difficult to comprehend, and it will only continue to grow.
An Open Source Approach.
The best way to tackle the protocol overload problem is to apply an open-source approach to standardize far edge IIoT protocols.
The mathematical concept known as the law of large numbers offers a framework for how developers can collate and aggregate all the data collected at the edge in a meaningful way to create business value. The law of large numbers states that as a sample size grows, its mean gets closer to the average of the whole population and, by extension, the true value. If you have a sufficiently large set of numbers, they will ultimately coalesce into the true value.
This law demonstrates why an open-source approach can solve the protocol problem. My thesis is that having hundreds of people contribute to protocol coverage in an open source way will enable us to address 90% of the protocols in the IIoT space. With a problem of such a massive scale, the solution also needs the ability to scale.
Applying the law of large numbers means organizations don’t need to change the work they’ve already done, no matter how many protocols they’re running. They can look upstream to where a project was forked to see if someone had already contributed to that protocol in the Akri community, a Kubernetes Resource Interface.
Similarly, if a developer needs support for an obscure protocol, they can go to the creators of said protocol and get the source code. Then, software engineers can incorporate it into the product and support that device.
Top protocols that are already aggregating are a sign that the law of large numbers is true in this case. However, we are still in the early stages. It’s up to developers and the open source community to bring edge innovation forward.
Rallying the Open Source Community
The open source community has been called on many times to help solve a common industry problem or pain point – in fact, that’s how Linux was born. Now, I’d like to call on the open source community to tackle this IIoT protocol problem.
One of the best ways to put this theory to the test would be greater participation from developers in the Akri community. Akri manages all protocols as Kubernetes objects, creating a standard way to manage devices no matter the protocol. Upstream participation would help improve existing protocols and broaden coverage for new ones.
To formally standardize IIoT protocols, we will need the involvement of the big companies that usually create them. Business leaders who understand the benefits of a common protocol driven by open source should push their teams to adopt such a model and become a catalyst for the change and open source adoption in the IIoT space. Just consider the impact of Project Sylva on the telco industry. This framework, created with some of the biggest carriers and vendors in Europe, has reduced complexity and accelerated the cloudification of telco networks.
The protocol issue is a deeply complex one, but the law of large numbers tells us that by putting the open source community on the case, it can not only be solved but spark even more opportunities and innovation at the edge.
##
ABOUT THE AUTHOR
Keith Basil leads product management, strategy, and marketing for the SUSE Edge business unit. He serves as a product champion, innovation and growth catalyst, and strategic leader driving cross-functional alignment, efficiency, and results.
