IoT at the Network Edge

What is the Internet of Things (IoT)? This is the question that everyone is asking nowadays. However, the answers are usually overly complicated and only produce more confusion.The simplest explanation is that IoT is an idea of connecting “anything” to the internet or to other connected devices with the simple concept of collecting information and sharing it.

‍

The problem that we were addressing

Rapid growth of IoT devices in the manufacturing process led to the point that we can no longer efficiently interact with every single device and rely on a centralized Cloud platform to process the data.

For environments where a remote piece of equipment must report to a supervisory system when an operational discrepancy is reported, the lack of local processing to extract the valuable data is crucial and leads to latency issues and higher costs.

We needed orchestration layer in-between to boost the performance and reliability of this solution. This is where Edge computing comes in by processing data closer to where it’s been produced.

‍

The approach we took

The solution we were seeking needed to be scalable, versatile and reliable. When architecting this we mixed Cloud computing and Edge computing in order to boost the performance and to provide all foundational elements for any successful IoT solution.

To deploy on various distributions and turn any device to handle processing functions natively we chose to use AWS Greengrass. Greengrass extends the AWS IoT platform, making the process of securely enrolling and managing devices into the ecosystem much simpler and allows us to adjust and ensure the improvement in performance.

By leveraging the network edge we were able to deploy this solution on top of industrial routers. The result procures platforms that locally execute our business logic and transfer only the computed data to the Cloud using 4G connection. Data caching was used to assure that no data will be lost when there is no connectivity to the Cloud.

The end result delivers features such as machine-to-machine communication, shadow implementation, various protocol communications with industrial equipment and remotely managed updates.

Furthermore, with the compute layer of the solution being delivered through Lambda and Edge computing being the next big thing in the IoT industry, more and more use cases will rely on the edge layer.