High-speed communication and Edge Computing are closely intertwined – you can’t have one without the other. Ethernet ensures the fastest possible transfer of device data to the edge node. While Edge Computing enables the performance goals of the 5G telecommunications standard to be achieved in the first place.
Even if it might be theoretically possible to equip every networked device with its own intelligence and sufficient processing capacity. And due to the rapid development of semiconductor technology, it still isn’t worthwhile. Firstly, the cost of doing so would be prohibitive. Secondly, many applications require data from more than just one source. One example is predictive maintenance in an industrial setting. An analytical program can only predict a component failure once information is supplied by as many sensors as possible. And installed at various points on the machine.
This is why the data are collected and processed in a node – or a mini data centre. A network architecture of this kind is frequently referred to as Fog Computing. While others still consider it to be Edge Computing. Regardless, many of the applications that run here require high-performance communication systems. Using which the data from the sensors, actuators or devices can be transferred to the mini data centre. And then back again in real time.
High-speed communication with Ethernet
Over the past few years, one communication system has been able to establish itself in many areas. Whether in factories, cars or buildings: Ethernet. When compared with traditional bus systems, Ethernet boasts higher speeds, improved handling of large data volumes and lower costs. Because of the technology’s high energy efficiency. Not to mention its flexible and economical bandwidth options. Nowadays, Ethernet protocols enable high-speed communication with speeds of up to 400 Gbit/s through the use of optical-fibre cables. For industrial applications, real-time Ethernet and industrial Ethernet options have been developed that operate with the lowest possible latency.
In the words of John D’Ambrosia, Chairman of the industrial consortium Ethernet Alliance: “Ethernet encompasses a vibrant ecosystem of technologies integrating seamlessly to deliver the robust connectivity demanded by emerging markets and applications. And even as networks grow larger and faster, Ethernet is successfully meeting the multi-vendor challenges of these networks head-on. Ethernet has not only horsepower needed to support next-generation networks, but is mature enough to do so reliably and cost-effectively.” Accordingly, the consortium assumes that high-speed communication speeds of up to 800 Gbit/s. Or even 1.6 Tbit/s will be possible in a few years.
A mobile-communications standard for the IoT
The connection of the devices and nodes networked via Ethernet with the cloud will then be established using broadband cables. Or – in the case of mobile applications – via cellular infrastructure. The cloud will provide highly flexible, scalable resources such as storage or (non-time-critical) computing power in central data centres. However, the sheer number of future networked devices is already pushing today’s data transfer systems to their limits. At least as far as the wireless technologies required for mobile applications are concerned. After all, there is the fastest mobile-communications network available today – LTE or 4G. It can only support 2,000 active devices within an area of one square kilometre. However, the continuous growth in the number of networked devices in the IoT will necessitate considerably more active connections.
The solution lies in the 5G standard, which can support up to 100,000 active devices per square kilometre. With that figure even rising up to 1 million in the future. Yet even with a data transfer rate of up to 10 Gbit/s, 5G can only work in combination with Edge Computing. Firstly, with the gigantic number of networked devices that there is anticipated to be in the future, the mobile-communications network would simply be overloaded if all the data involved were sent to the cloud to be processed. Secondly, applications such as autonomous driving or Industry 4.0 require latency times of less than 10 milliseconds, with some even as little as 1 millisecond. In order to achieve these response times, the transfer distances for 5G cannot be too far either.
A mini data centre in every cell tower
That is why 5G providers position data centres at the edge of their mobile-communications networks; more or less right at the 5G antenna itself. This enables them to host or provide a range of applications and services that benefit from low latency. At the same time, they can reduce the volume of data traffic that needs to be fed back to the core network. This reduces the cost of the data transfer – in short, edge computing saves mobile-communications providers money. It also leads to an improvement in the overall mobile user experience for consumers: the low latency results in a palpable acceleration of the response within mobile-communications networks. For instance, if TV shows and films were to be temporarily buffered at the edge, video streaming on mobile devices would begin almost immediately.
To put it simply, 5G would not be able to achieve the targets of very low latency and massive broadband without edge computing. “Edge computing only emerged on the wireless-industry stage several years ago, yet its significance is already great,” says Iain Gillott, President and founder of specialist consulting company iGR. “We believe that edge computing will ultimately be essential in realising the full promise of 5G.”