Edge Processing: What is it and Why Does it Matter for the Rail Industry?

The technology advances disrupting transportation depend heavily on edge processing—a topic that is generally not well understood, and one I am asked about frequently by my rail industry colleagues.  I often recommend Daniel Newman’s recent Forbes article, Moving To The Edge: Evolving IoT Data Storage In 2018which does a great job of explaining edge computing in layman’s terms.  Since sharing it, however, I have received several questions regarding specific applications to rail.


Edge computing, or processing, is a key innovation in IoT, and one the rail industry needs to understand and adopt quickly.  According to Gartner, 10% of enterprise generated data is created and processed on the edge today, and this will increase to 50% over the next 4 years. Given the fact that railroads have thousands of remote, mobile assets, the industry should expect an even greater percentage of enterprise data generated and processed on the edge.


On a practical level, edge processing provides the ability to handle data that is too big to send off to the cloud wholesale.  It also allows you to use data to take action in real time. This means rail operators can detect and manage critical issues that need immediate intervention, such as when a train that is exceeding speed restrictions.  It allows rail companies to send real-time alerts or warnings to the crew operating a train, or to automate braking if the train does not slow when it should.




Let’s start with what it isn’t…


Connectivity is often mistaken for IoT and edge processing. Don’t confuse an onboard device with an Ethernet cable with edge processing.  For edge processing, onboard technology must have full-fledged operating systems capable of running complex algorithms.


Edge processing occurs closest to the data source or point of data collection, such as within a mobile device—in our case, the locomotive or rail vehicle. It allows data captured via IoT solutions, such as remote monitoring, to be processed closer to where it is created, instead of sending it across long routes to data centers or the cloud.


Analyzing video and other high-resolution data on the edge enables you to identify and send only the important data to the cloud.  Miles of uneventful video taken en route to a train’s destination can be eliminated, while video of a near-miss at a crossing, which is critical for compliance and improving safety, is transmitted.


Some processing can be handled more reliably on the edge, because it doesn’t require continuous connectivity to the network.  You can instantly analyze data to take real-time action based on algorithms designed to meet your parameters. When your trains travel in areas where networks are less reliable, your crews still get the information and support they need.


This is essential when real-time processing is critical, as in the example of delivering real-time over-speed warnings to locomotive crews.  If a train is barreling down the track at unsafe speeds, even sub-second cloud-processing response times are too long.




Unlike consumer mobile devices, locomotives and rail vehicles are designed and manufactured to be in service for many years.  This means rail companies must select a platform that will keep up with the pace of changes in data and software over the life of their assets—how is that possible when the nature of those changes is unknown?


Good architecture designs for the unknown.  The onboard technology must be rugged computers designed for long life and high reliability, which can be remotely updated on a frequent basis.  Flexibility and extensibility must be built into the fundamental design of the overall platform to ensure that the system can handle new problems over time and can reconfigure what gets processed on the edge vs the cloud as problems and data evolve.




Edge computing transforms typical locomotive and rail vehicle remote monitoring into an engine that drives critical action as the data is being collected.  How much better would you operate if you could take the insights from your system and automate actions aligned with your well-informed decisions? This is the promise of IoT.


Using video data to remotely inspect rail infrastructure, we see clients reducing costs associated with manual track inspection and significantly increasing efficiency even over what can be achieved using drones and other remote alternatives.


These types of efficiencies, enabled through IoT and edge processing, are no longer optional for rail companies when innovators like Elon Musk and Tesla announce their intention to beat rail efficiencies and disrupt the industry with their electric semi-trucks.


Originally published on LinkedIn