After "Cloud Computing", a new term "Fog Computing" has been added in recent years. It mainly comes from the phrase "fog is a cloud closer to the ground".
For the development of IoT technology and applications, more and more intelligent networking devices are connected to the network. Under the basic model of cloud computing, most of them have only very weak computing power. Therefore, related computing resources, such as storage or computing power, must be accessed from the cloud through the network. With the rapid growth of demand, the burden on the cloud is getting heavier and heavier. After all, it is a centralized resource. Secondly, the cloud is relatively far from the end, and these smart devices have a long delay when accessing resources on the cloud.
Fog computing is designed for data-intensive, high-performance computing, and high-risk environments. Fog is an emerging distributed architecture that bridges the cloud and the devices connected to it, without the need to establish a permanent cloud connection between the site and the factory. By selectively transferring computing, storage, communication, and control, fog computing can make decisions close to IoT sensors and actuators (this is data generation and use). It is a useful supplement to cloud computing, not a complete replacement so that IIoT can be used efficiently, economically, safely, and constructively in a manufacturing environment.
Fog is sometimes called edge computing, but there are key differences between them. Fog is a superset of edge functions. The fog architecture combines resources and data sources with a hierarchical structure that resides on north-south edge devices (cloud to sensors) and east-west edge devices (function-to-function or point-to-point) for maximum efficiency. Edge computing is often limited to a small number of north and south layers, usually related to simple protocol gateway functions.
Therefore, under these considerations, there is fog-end computing, which means that computing resources are decentralized to some extent and deployed closer to users.
Therefore, fog computing is not a substitute for cloud computing, but as an extension of cloud computing.
Basically, the combination of the two is to hierarchize the allocation of computing resources. The top layer is the cloud, the middle layer is the fog, and the bottom layer is the client-side connected device. Therefore, under this model, a certain percentage of computing resources that were originally concentrated in the cloud will be reduced to the fog end. When the device wants to access resources, it will be accessed as close as possible to the fog end.
In this way, because the fog end is closer to the ground, the access speed is faster. Secondly, even if frequent and large amounts of communication are required, a large amount of network traffic will only be scattered between the ground and the fog, and will not communicate. Throwing into the cloud reduces the burden on the cloud.
Factories can make full use of the data flow of the fog node layer to make the connection between factories better. Fog nodes located at a lower level in the overall structure, such as a single computer, can be directly connected to local sensors and actuators, so as to be able to analyze data in time and explain abnormal operating conditions. If it has been authorized, it can also respond and compensate for problems or solve problems autonomously. In addition, fog nodes can also send appropriate service requests for higher-level fog hierarchies to providers with better technical resources, machine learning capabilities, or maintenance services.
If the operating conditions require real-time decision-making, such as shutting down the equipment before it is damaged or adjusting key process parameters, the fog node can provide millisecond delay analysis and operation. Manufacturers do not have to use cloud data center routing to implement this real-time decision. This helps avoid potential delay issues, queue delays, or network/server downtime, and these delays can cause industrial accidents, reduce production efficiency, or product quality.
In the factory, the fog nodes located at a higher level can obtain a broader perspective on industrial processes. They can add more functions, such as the visualization of production line operations, monitoring the status of malfunctioning machines, adjusting production parameters, modifying production plans, ordering supplies, and sending alerts to the right people.
Fog computing can help IIoT and smart factories bring various benefits, including productivity, product quality, and safety. IIoT can provide a technical route for clean and green manufacturing. As a result, the manufacturing industry will achieve unprecedented customer-level collaboration and achieve mass customization and large-scale personalized customization. The potential opportunities to take full advantage of all aspects of the Smart Factory are endless.