Edge computing is gradually and profoundly changing the way we manage data. How does it work? Process data as close as possible to the end user or the connected objects that generate it., without having to transfer them to the cloud. The result: a faster, cheaper and more reliable system. We tell you all about it.
Traditionally, data is stored and processed at both ends of the network :
- either as close to the user as possibleon our computers, smartphones or on the company's local servers,
- either a long way from the user, on remote servers within gigantic data centers, in the cloud.
Edge computing defines a third way.
Definition of edge computing
As its name suggests, edge computing involves processing data in a distributed way, at the edge of the network, rather than centrally in the cloud. It uses mini data centres to store and process data as close as possible to connected objects, end users and businesses. The result: better quality of service by improving latency, throughput, energy consumption and environmental efficiency. Once processed, some of the data can be sent to the cloud for other types of non-urgent processing.
So edge computing is nothing new in itself. Distributed computing has been around for a long time, but its adoption and applications are growing fast thanks to the boom in connected objects, which are putting a strain on the bandwidth of networks linked to the cloud.
Ideal for processing data in real time
The cloud is facing some difficult challenges: processing more and more data, and processing it faster and faster. For certain applications (industry, energy production, medicine, road traffic management, etc.), processing times and reaction times are critical. Data must be processed in near real time.
In the industry
Red Hat, an infrastructure solutions company, provides a good example: in a modern manufacturing plant, some 2,000 Internet of Things (IoT) sensors monitor production. These connected sensors generate around 2,200 terabytes of data per month. A mass of data that can be processed more quickly just a few kilometres from the factory than in a datacentre on another continent.
In logistics and mobility
Another example is connected vehicles. Buses, trains, trams and lorries are equipped with GPS chips and sensors to monitor traffic in real time and analyse passenger flows. The on-board computer provides real-time information on traffic conditions (accidents, roadworks, blocked roads, etc.). This enables delivery drivers to find the fastest routes according to their delivery plans. It's hard to imagine waiting 30 minutes for the route to be updated and then getting stuck in traffic because the data took too long to be sent and processed.
The number and uses of connected objects and the Internet of Things (IoT) are exploding: 13 billion worldwide by 2022, almost 30 billion by 2030, according to Statista. In France alone, Ademe and Arcep put the number at 244 million. They include our mobile phones and laptops, as well as robots, cameras and sensors used in industry and smart cities. The development of 5G will further accelerate this phenomenon.
And to process all this data, edge computing is the obvious solution. According to Research & Markets, edge computing is growing by 17 % a year.
By 2025, Gartner estimates that more than 50 % of the data managed by businesses will be processed outside the cloud.
The benefits of edge computing
Reduced communication latency
Processing information as close as possible to its source considerably reduces the time needed to transfer data and react accordingly. For example, in the event of a breakdown on a production line, this will enable IIoT (Industrial Internet of Things) sensors to transmit the data and its analysis as quickly as possible.
Higher performance
By transmitting less data in the cloud, the data will be processed faster with a higher throughput. The company will also be able to cache data in a CDN (content delivery network) integrated into the edge network. This will enable content, particularly video, to be distributed more quickly.
Reduced infrastructure costs
By reducing bandwidth and data centre resources, edge computing delivers substantial savings.
Greater scalability
Distributed edge computing makes it possible to modulate the resources needed to process and store data.
Environmental efficiency
By saving IT resources and reducing the volume of data exchanged, edge computing emits less CO2 and will have less impact on the environment.
Is edge computing replacing the cloud?
Edge computing also has its drawbacks. The first is having to use service providers who will offer this edge service. Depending on the company's location, the choice is not as wide as for cloud providers.
The other disadvantage concerns cyber security. The more information is processed at the periphery, the more complex the architecture becomes. Equipment and service providers are more likely to be heterogeneous. The hardware and processes may be different, with several suppliers... which increases the vulnerability architecture.
So edge computing is not a replacement for the cloud. It complements it by handling data with high processing speed requirements. Other types of data, whether less critical or more massive, can continue to be processed in the cloud. So the cloud is not about to disappear.
