Edge Computing: The Cloud, the Fog and the Edge
April 23, 2017
April 23, 2017
With the concept of Internet of Things and automation erupting all around us, connecting billions of devices worldwide, the role of powerful networks has never been more important. At the heart of these networks is the cloud which has become the central focal point connecting all applications and devices together.
But is the cloud enough? – Could edge computing be the answer?
The heart of a connected network is the cloud. The cloud processes data, sends out information to devices, includes all the algorithms and software to run all applications and acts as a central management point for the whole network. There is no doubt that the cloud is the future of automated networks, it is the most optimized way to run a network – most industries and companies are well aware of this and now setup their networks round a cloud.
In recent years new concepts were conceived in an aim to enhance the optimization and operational capacity of cloud-based networks. These concepts add processing units closer to the devices as mediators at the physical edge of the network; this is known as Edge Computing, and it redefines the way networks operate.
So what’s fog computing and what’s mist?
Fog and mist are conceptual network levels that operate close to the base level of the network. So the cloud overlooks and controls the entire network, the fog connects end devices and the mist adds micro-controllers between certain devices. It’s easy to visualize the cloud hanging in the sky, as the fog is a smaller cloud spread out over the ground, and the mist hangs lightly between the different elements at the ground level.
Depending on the definition and network complexity, fog computing and edge computing can be considered synonymous – seeing as how both concepts define the same area of the network. In larger networks with more connected devices of varying types, edge computing could be defined at a higher level connecting a number of fog devices, which in turn connect a number of end devices each.
The benefits of edge devices are immense, but what truly makes this concept groundbreaking are the many possibilities that have recently opened up thanks to advances in this field.
One huge ability of the edge is de-cluttering large networks with many levels. As we’ve seen the different edge levels operate at the physical tier of the network, between the end deceives, optimizing the communication with the cloud – and lowering the strain on the network. So, there is a clear benefit to IoT and automotive infrastructures. For example a fleet of buses needs to upload, access and download information from the cloud in real-time, updating travel times, vehicle status, route selection, capacity, and so on. With, say, 150 buses across a small city – the strain on the network would be incredible and there will be widespread latency issues. Adding edge devices on every bus, bus stop and depot can act as mediators – processing the raw data and sending only a small selection of crucial information to the cloud, is exactly what such a network would need.
The “de-cluttering” effect of the edge concept has another great benefit, and it has to do with media.
Media files are much heavier than simple data, naturally with higher resolutions the bandwidth can become very strained. Imagine a company trying to develop an AR device for construction, safety, medical or repair teams that displays media in real-time, at a high resolution and projected onto an eye piece for each team member. This would allow them all to view a real-time plan of the work they need to carry out, all controlled by the team supervisor and later accessed for reporting and training. Sounds nice? Well without edge devices mediating all this intense data, non of this would be possible, not in real-time, and not in a high enough quality to work with.
In order to achieve an optimized network connecting hundreds, thousands or even millions of devices, sensors, data terminals, management modules, and so on; edge devices (including fog and mist devices) must be powerful enough to handle the network load. These devices must be at industrial grade with a relatively long life span, in order to withstand the harsher conditions at the network edge. Depending on the type of network, the edge device could be at power control boxes along the street, attached to a smart vehicle’s engine, or connected to cell towers at remote locations not easily accessible for on-site maintenance.
So, whereas the cloud management and network servers sit in a controlled environment with access to on-site maintenance, modification and upgrades – the edge devices must be much more autonomous, strong enough to withstand different conditions, and maintained, modified and upgraded remotely for the most part.
To answer the growing needs of Edge Computing networks, SolidRun has developed the ClearFog line of single board computers. These robust SBCs address all the needs of the modern edge-based network, providing powerful processing and connectivity, built into an industrial grade board that can be placed where needed along the network.
We believe in modularity and flexibility, allowing developers absolute freedom in creating a tailor made network that fits to the specific needs of their network.
Have any thoughts about Edge Computing? comment below.
Also, take a look at what SolidRun has to offer in embedded systems solutions.