Air gap solution lets farmers deploy machine learning without internet connectivity

Making sure your agricultural equipment is using all the latest and greatest machine learning models can be very difficult. Connectivity can be sparse and expensive if you are using satellites, says Jason McCampbell, Director of Architecture at Wallaroo.

His company therefore launched Air Gap Edge Deploy functionality to make it easy for enterprises to deploy and manage machine learning models at the edge in environments with no IP connectivity. Think of oil derricks, gas pipelines, and transmission lines for Energy & Utilities; and autonomous equipment in Smart Manufacturing and Smart Agriculture.

We talked to Jason McCampbell about the air gap solution and how it could benefit farmers.

Why was Air Gap Edge Deploy functionality developed?

“There are many reasons why enterprises (not just farmers) could be looking to deploy machine learning via air gap. For example, equipment could be far from internet connectivity at the edge, such as in oil derricks, gas pipelines, or in agriculture, perhaps a combine out in a rural area far from a cell tower. Additionally, the increase in cybercrimes has led to some companies exploring the option of air-gapping to keep their systems secure. By isolating their networks from external ones, they are able to prevent the vulnerabilities that come with having these connections like data breaches and ransomware attacks that can cost billions in losses.”

In a Future Farming Field Trials podcast earlier this year, Jesse Hirsh, technology strategist, futurist and livestock farmer detailed how people living in rural areas should not wait for large communication companies and governments to bring internet signals to their door.
Check out this Field Trials podcast: – Solving the poor internet problem

How does Air Gap Edge Deploy functionality work?

“You can train a machine learning model anywhere, in the cloud or on premise. But from there you will need a remote connection (for example, to the actual farm equipment where you want to deploy the model. We provide more details on how it works here.”

Why is it relevant for crop growers?

“Agriculture equipment is generating more data than ever. The CTO of John Deere in an interview with The Verge said their farm equipment had basically become “mobile sensor suites that have computational capability.” Combining this sensor data with AI enables crop growers to use just the right amount of water, fertilizer, pesticides, etc for each individual plant. Additionally, crop growers are relying on robots more for picking different crops using computer vision and so on.

Essentially what data plus AI gets you as a farmer is a greater yield while lowering your input costs. But making sure your equipment is using all the latest and greatest machine learning models is very difficult. And you also have to think about the flow of data back to the data scientists who trained the model. All this sensor data can be gigabytes or even terabytes of data per day.

Connectivity is sparse and can be expensive if you are using satellites. So for the owner or servicer of equipment, it is much more cost effective to use something like this air gap solution to deploy the machine learning model into the equipment as well as to take out the production data so you can make sure your models are still being accurate and performing well. More information on what this flow of data in and out looks like here.”

What are the costs for a grower using this air gap solution?

“As you see a greater shift into equipment as a service (EaaS), more of these costs will be shouldered by the owners and servicers of the farming equipment who then sell it through to farmers as part of a larger smart agriculture services. So a farmer will pay for AI services, which will include edge machine learning as part of it.”

And what are the financial benefits for the growers?

“The adoption of machine learning into agricultural production has become a necessity given the needs to increase food production while balancing sustainability. ML has already started making an impact, providing insights that help increase productivity, using less water, fertilizer, pesticides, etc. The data processing abilities granted by ML have led to early adopters expanding the possibilities of crop growth into areas previously unsuitable and increasing overall yield.”

Can you give a practical example?

“We already have examples of farm equipment automatically adjusting water, nutrient, and other chemicals used down to the plant level by using sensor data combined with AI. In terms of our own air gap solution, we do not have examples to share but we do in other industries, especially manufacturing. In fact, we are working with the US Space Force around edge machine learning deployment for their satellite fleet.”