3.1.11 Fog computing model
Cloud computing has solved many problems of the traditional client–server model. Cloud computing may not be the best option for delay-sensitive applications that require an immediate, local response.
The emerging wave of the IoT requires mobility support and geographical distribution, in addition to location-awareness and minimised delay. Devices in the IoT will require real-time data and quality of service mechanisms. The IoT encompasses an almost limitless number of IP-enabled devices that can monitor or measure nearly anything. However, the one thing these devices have in common is that they are distributed throughout the world.
One of the most significant challenges this presents is creating links between these devices and the data centres where data can be analysed, as shown in the figure. These devices can produce huge amounts of data. For example, in just 30 minutes a jet engine may produce 10 terabytes of data about its performance and condition. It would be inefficient to deliver all the data from IoT devices into the cloud to be analyzed and then forward decisions back to the edge. Instead, some of the analysis work should take place at the edge, for example, on industrial-strength Cisco routers built to work in the field.
Fog computing creates a distributed computing infrastructure closer to the network edge that carries out easier tasks that require a quick response. It reduces the data burden on networks. It enhances resiliency by allowing IoT devices to operate when network connections are lost. It also enhances security by keeping sensitive data from being transported beyond the edge where it is needed.