Fog computing may be IoT’s computational model
Fog computing and fog networking could fill the latency and range gap in the internet of things (IoT.) For the last couple of years, researchers have been reporting on developments in fog’s role in completing IoT’s ubiquitous connectivity. It is similar to cloud computing architectures but it brings the cloud to the edge to meet the different demands of IoT.
Source: Steven Max Patterson
The underlying concept is the cloud for some real-time IoT services could be too slow because the quality of service (QoS) specifications for the IoT application exceeds the cloud’s QoS. The solution is to move the cloud out into the network.
Low latency and QoS is important in IoT use cases like self-driving vehicles and controlling robots and other control applications that require minimum latency to synchronize, supervise, control and initiate machine actions. Range is important when connecting devices over long distances where hubs and gateways are not locally available.
The telcos have staked out this market and have begun a soft sell that machine to machine (M2M) communications and 5G is the solution for IoT connectivity while they prototype early 5G and more mature M2M equipment in their test beds.
A first reaction to the telcos might be one of suspicion that this is just another sales pitch for a new type of IoT messaging service; however, the telcos have a point. Unless the laws of physics change, the low latency, and long range for some applications will need powerful grid-connected shared radio networks and computational resources operated by the telcos that already own these radio networks and backhaul facilities.
The planning has already begun. A year and a half ago, telco giants, AT&T, Deutsche Telekom, EE, SK Telecom, and Verizon joined the Open Compute Project (OCP) to form the Telco project with the intent of creating open computational hardware and software designs for edge computing. The OCP Telco project can be categorized as a software defined network (SDN) initiative and fog networking and computing, a subcategory under SDN for IoT.
Fog computing and fog networking are often used interchangeably. Both terms describe a very large number of interconnected heterogeneous and decentralized devices that can communicate and cooperate with each other over the network without interacting with computational services at the edge.
Mobile Edge Computing (MEC) brings more computational power to the edge of the network extending the capability of the interconnected IoT network. IoT devices are hardware-constrained by cost, the computational capability of microcontrollers and storage capacity of memory. Computation exceeding the capability to predict the next local action of the device would take place at the edge. For instance, an IoT device connected to a traffic light may need to compute its updated prediction for changing from red to green based on a histogram of passing vehicle traffic too large to be stored in memory. The prediction would be computed at the edge on servers and sent to the device.
From bots to big data, cloud technologies spur today's innovation and tomorrow's bottom line. Our survey results reveal why.
An intelligent traffic model at the edge may trigger a change in traffic management computed on the aggregate of IoT devices to optimize the capacity of the smart city’s roadways and minimize travel time.
Some M2M networks are already in operation. Comcast recently completed the construction of a LoRa wireless network blanketing the Philadelphia metropolitan area designed for healthcare, public utilities, automotive and consumer electronics. LoRa is one of a number of low-power wide area networks (LPWAN) that operates at 0.3kbps to 50kbps at a range of up to 15 km.
Another M2M network Aeris uses existing cellular technology. SigFox, operating with a range up to 40 km is notable because its radio technology penetrates urban and industrial landscapes. The choice of these wide area IoT networks is a trade-off between signal strength, range, throughput, and power consumption.
When 5G arrives, these networks and ones like them will not be obsoleted because one network will not be fit based on these tradeoffs. Fog Radio Access Networks (F-RAN) are under discussion to consolidate these heterogeneous networks into a single network architecture with 5G even though they do not operate in the same bands to gain high spectral and operating and energy efficiency. The 5G networks are reported to use 90% less energy.
F-RAN would bring 5G MEC computing to these networks. The consolidation would also break down and group functions such as signal processing, precoding, control and user planes and computing improving overall utilization. A single telephone, a single computer or a single IoT device do not add much value. As the numbers of interconnected devices IoT devices increase, the network effect will increase their value in the same way it did for telephones and computers.
Fog computing will not replace cloud computing, and they are likely to be interconnected and operating different services best suiting the proximity to the device. Providing enough coverage using the many wide area IoT choices will increase the IoT network effect.