Edge Computing: how to bring intelligence and responsiveness closer to the source

With the development of 5G, you are going to hear more and more about it and yet, this is nothing new: theEdge computing. Behind this mysterious expression, hides a technology which comes to deport part of the computing power as close as possible to the field. This is the case with sensors in industry, for example, or those of smart cities, or more precisely for autonomous vehicles.

This process is also used in special cases, such as for the protection of the environment, such as hardware developed by Dell Technologies and Intel : this one is embarked on tourist ships to retrieve pictures of the reefs of the Great Barrier Reef in order to help her backup. In this case, Edge computing then makes it possible to collect the data autonomously and transmit them to servers scientists as quickly as possible.

But, beyond simply transmitting raw data directly to data centers to process them, Edge computing also makes it possible to analyze them directly. A practice that has intensified with the generalization of the use ofIntelligence artificial. Rather than simply observing and noting a situation collected from sensors, Edge computing “doped” with AI now predicts the future by identifying the risks of breakdowns and issuing alerts before those. -this do not happen and this, far from the data center.

Better yet, based on this prediction, the system can even adopt corrective measures automatically when possible. To achieve this, no need for the strike force of a data center whose connection is sometimes insecure or slow due to the distance. The main advantage of Edge computing is precisely that it makes several gateways Edge positioned as close as possible to their sources. This makes it possible to pool their resources in order to benefit from sufficient computing power to use this AI.

It’s not just the cloud in life

With such a network, we obtain a latency conducive to almost real-time management, whereas this is not the case when it is necessary to push data to the data centers and obtain instructions in response to its processing. Obviously, everything will depend on the type of data collected. Some do not require short response times. The connection to a powerful data center remains essential when it comes to collecting data allowing the regulation of background operations linked to the systems. But, when it comes to solving a concrete local situation, Edge computing brings more fluidity.

Here is for the theory. In practice, setting up such a system requires positioning a real small IT processing infrastructure as close to the field as possible. It must be autonomous, powerful and robust enough to withstand extreme temperatures, dust, humidity, shocks … In other words, it must work like proven military equipment.

Specialists, such as Dell Technologies, determine the combination between constraints of robustness, processing power, storage density, economy in energy and theautonomy for Edge units. And then, it is necessary that all this small world can work together with sometimes a multitude of protocols different communications, such as links ADSL, satellite, 4 or 5 G, or even links radios specific.

Under these conditions, the ideal conductor for getting data through quickly is the SD-Wan. Finally, we must not forget that given the sometimes improbable locations where these units are placed, it is necessary that the maintenance can be performed reliably without requiring direct intervention. And, in the event of a major failure, repair operations must be able to be carried out as quickly as possible and on site. As you can see, we have not finished hearing about the benefits of Edge computing.

Produced in partnership with Dell Technologies teams