The Internet of Things (IoT) poses a new challenge to IT. IoT components already generate enormous amounts of data that are more than a match for most corporate infrastructures. That explains the triumphant advance of cloud computing, which helps companies to achieve more agile business processes and paves the way for them to tackle digital transformation without having to make costly investments.
Much of the data that IoT applications generate is stored in the cloud. Most IoT devices are small, have a very low power consumption and perform narrowly defined functions. So the cloud is the perfect partner for the processing power that is required because it offers, in addition to computing capacity, storage options and security solutions. If information is required locally or at short notice, it is a different matter. That requires short latencies, which are not, as a rule, available in the cloud. Data generated in use cases of this kind mostly need to be processed extremely fast, preferably in fractions of a second and on the spot. Storage plays only a minor role (if any). So the computing power for these uses does not need to come from the depth of the network or from a cloud; it can comfortably be provided from the edges of a network. What used to be called edge computing is now known as fog computing. The term is considered to have been coined by Cisco as an extension of cloud computing. “The main task of fogging is positioning information near to the user,” Maher Abdelshkour explains in the network equipment vendor’s blog.
Distributed computing power for swift responses
Using this technology, data is not processed centrally in the network but at its edges. As technology has advanced in terms of rising chip performance and falling electricity consumption, small computers are now able to carry out the complex calculations required and are optimally suited for use in a fog computing scenario. Powerful servers are not required; smaller, distributed computers are used in the fog, serving as digital contacts for different devices, production facilities, streetlighting or vehicles and much more.
A use case for fog computing is, for example, Car2Car communication in the event of emergencies. If a vehicle initiates an emergency braking maneuver on the road, other vehicles within a certain radius must be notified to prevent rear-impact crashes. This situation calls for lightning-fast reactions, so it makes no sense to first send the acceleration data via the cloud to a data center for affected vehicles to be identified and notified. The time delay would be too significant. That is why the data processing and notification of nearby vehicles must be undertaken with the assistance of local computing power. For this task, fog computing is the right platform.
Cloud and fog make perfect partners
Fog computing will not supplant the cloud; experts see it more as a perfect partner. It can be visualized as an additional, virtualized layer between the terminal devices that generate the data and the cloud. It serves as what might be seen as a local decision-making and processing level that eases pressure on downstream facilities. This is the layer in which the rules for individual use scenarios are formalized, the rules on which the processing and provision of data are based. It reduces bandwidth requirement and optimizes response times so that fog computing makes applications possible that would not be possible with the cloud alone because the bandwidth required does not exist – and because these applications generate data that is only of relevance at a specific location and in the context of local conditions and does not require central processing.
The trend toward fog computing is only just beginning and is closely coupled to the IoT’s increasing market penetration. For an efficient and inexpensive Internet of Things, fog computing is set to become a key criterion as more IoT devices enter into our working lives and our private surroundings.