When we thinking edge computing we typically think of an endpoint device such as a smartphone or an IoT gateway that connects to an edge or cloud infrastructure. We generally assume some form of data or computational offload or distribution between a cloud or edge server to an endpoint client. What about endpoint devices communicating with each other and sharing resources at the edge? It’s another paradigm of edge computing that doesn’t get talked about as much but is becoming an increasingly interesting and important architecture.
What is peer-to-peer edge computing?
Peer-to-peer (P2P) edge computing is an archetype of distributed computing architecture that shares functions and/or workloads across peer devices or compute nodes in a network. At the mentioning of peer-to-peer applications, you might think of Napster, it is not what is meant mean by peer-to-peer edge computing.
Peer-to-peer edge computing occurs in an edge environment such as a home, or a factory, or campus, or the intersection of a street. The idea is to pool compute and data resources that are in proximity to where an edge scenario takes place in your business, usually in a local environment.
Peer-to-peer edge computing architectures and technologies enable a collection or swarm of intelligent connected devices to share their compute resources and data to perform augmented and/or collaborative functions without tapping into edge infrastructure or tromboning data or remote functions requests to the cloud. These distributed edge application designs can result in novel mesh intelligence and coordinated functions and capabilities. Because many of these endpoint devices are local, latency is less of an issue of physics.
What are good examples of peer-to-peer edge computing today?
There are a number of forms of peer-to-peer edge computing that exist today or are emerging.
You might have heard of swarm intelligence. This is an AI concept where a number of agents interact and coordinate their actions much like what we witness with ant colonies, flock of birds and such. A peer-to-peer edge application that is often referenced in smart transportation is the idea of platooning where a group of vehicles coordinate their position on a road to increase capacity, fuel efficiency, and safety.
One of the exciting things about 5G that has yet to gain mainstream traction is CV2X or Cellular Vehicle-to-Everything. It is a technology that was originally introduced in 4G to enabled vehicle to vehicle communications of all varieties to usher in the era of autonomous vehicles and smart transportation. CV2X continues to be developed as part of 5G and is expected to be fixture of mobile wireless technology standards for the foreseeable future.
Some peer-to-peer edge computing scenarios being piloted today include the smart intersection. Vehicles communicate with each other, with the traffic management system, pedestrians, and other elements on the road and its proximity in sharing contextual information. The idea is that this real-time contextual information will improve traffic efficiency, and pedestrian and vehicular safety.
Key Takeaways for the C-Suite
As we see more intelligent endpoint devices populate the “edge” we can expect peer-to-peer edge computing models to make sense, especially for applications that rely on time-sensitive synchronized actions by a population of coordinated devices. Most if not all of these applications will execute simple functions.
The technologies are already here to make peer-to-peer edge computing happen. 5G CV2X and extensions of this technology beyond vehicles and city infrastructures will inevitably enable a whole range of peer-to-peer applications in industrial settings with robots. Time and economics will tell when peer-to-peer edge computing becomes mainstream.
Watch the webcast by clicking the Accelerated Economy logo below and subscribe to the Cutting Edge column.