Pardon the pun, but companies in many different industries that want to get an edge on their competitors — or avoid being left behind by them — need to embrace the next step in the evolution of digital technology: edge computing. So, what is edge computing?
To put it simply, edge computing refers to digital data processing that is done in the same place or very near to where the data is collected. It gets its name from the fact that an “edge” computer is located at a great distance from a central computer (or the cloud) and handles data only for one or a limited number of devices or systems generating that data. This is in contrast to a centralized computer that collects and processes data from many sources, all of which are connected to the system by two-way wireless connections in a hub-and-spoke array.
To understand this difference, think of a typical telematics system. Sensors on a vehicle transmit data to a central computer, where it’s processed for information that is transmitted back to the vehicle. In this arrangement, the vehicle is somewhat like our earlier “dumb” desk top terminals from the ‘60s and ‘70s, which had no computing power of their own but enabled users to tap into one central, shared computer.
Using the same analogy, an edge computer is a bit like a laptop computer that can also be linked into a computer network. The edge computer can do some of its own processing for the user, while it can also send data and processed information to a central computer. In this way, an edge computer achieves a measure of independence from the master computer.
Minimizes processing time
Edge computing solves two growing and inter-related problems with this feedback loop. One is lag time, which IT professionals call “latency”. It takes time to complete the circular flow of data from a sensor to a central computer and back again, which brings us to the second problem: data volume. So much data can be generated that it taxes the capacity of the transmission conduit – its bandwidth — to send it quickly.
It’s a problem that video gamers experience from time to time, with the massive volume of graphics data occasionally causing games to stutter like a PowerPoint presentation momentarily freezing or even overloading the game to the point that it crashes. When it comes to games, that kind of system failure is annoying, but when it comes to an autonomous vehicle (AV for short) – the biggest projected use for edge computing – the lag can be fatal.
To avoid accidents, AVs will have to make nearly instantaneous decisions based on huge amounts of data, including the speed of the vehicle, the location and speed of other vehicles, the presence of obstacles and pedestrians, the configuration of the road, the road surface, and weather conditions, just to name a few. An on-board computer, powered by artificial intelligence (AI), will be needed to bring down the data processing time to crash-avoiding intervals.
Five uses for AVs
Blair Felter is the director of marketing at vXchange, a data-as-a-service provider based in Philadelphia. She writes that edge computing will be critical to the expected transition to fully self-driving cars and trucks in five different ways:
Data management. AVs will generate a staggering amount of data. In a single day of driving, an AV is estimated to generate 30 terabytes of data, much of it unstructured. Multiply that by tens of millions of vehicles (there are 250 million on U.S. roads alone) and it’s clear that wireless bandwidth will be severely challenged to avoid critical latency. Edge computing, Felter explains, will provide the extra computing power necessary for split-second analysis and vehicle response.
Vehicle to vehicle communication. To achieve maximum safety, each AV will need to respond not only to its own sensors, but data and information generated by other vehicles on the road. Examples include real-time weather, road construction, and traffic conditions during rush hour. Edge computers will enable AVs to send and receive that information without having to be connected to distant cloud servers. This will also be achieved via vehicle to vehicle communication.
Processing speed. On the road, extra milliseconds in data processing can make all the difference between a safe journey and a collision. Self-driving cars need to react immediately to changes in conditions, and the stakes are too high to allow a vehicle’s control system to suffer from computational lag. Edge computers, located in vehicles or in edge computing facilities in high-traffic areas and places with limited bandwdith, can minimize computer latency.
Smart city integration. For AVs to reach their full potential, Felter writes, high-traffic urban areas will need to provide vehicles with a wealth of information, including everything from road conditions to real-time reports on traffic congestion. The information will come from sensors installed throughout those urban centers linked to an edge computing network.
The expansion of fog computing. To provide further data processing power, between the cloud and edge computing a new intermediate realm has arisen called “fog” computing. This level involves the creation of micro data centers at the base of cellular towers to store and process data closer to where it’s collected. These centers decide which data needs to be relayed back to the network’s central cloud server and which can be processed locally. Serving gatekeepers for data traffic, fog data centers can communicate faster with local devices and improve overall network performance.
I’m in the automotive fleet service industry, so AVs are of special interest to people like me. But edge computing has potential application in a wide number of industries, like manufacturing, healthcare, finance, retailing, and security.
Investment in edge computing is exploding, testimony to the fact that businesses understand what it means for future competitiveness. Here’s the outlook from one market research firm, Grand View Research:
The global edge computing market size is projected to reach USD 3.24 billion by 2025, …expanding at a phenomenal [compound annual growth rate] of 41.0% …. [The] need for advanced technologies is stimulating the volume of IoT [Internet of Things] data. Vast amounts of data created by IoT devices may cause delays and latency. Edge computing solutions help enhance the data processing power, which further aid in avoiding delays.
Business leaders need to start asking themselves and their senior staff now whether and how edge computing could benefit their organizations. It’s a race you don’t want to lose.