The popular “cloud computing” is to put large amounts of data in the “cloud” for calculation or storage. In this way, the current computer or mobile phone storage capacity is not enough, or the computing speed is not fast enough, of course, it also brings many other benefits.
The core of this so-called “cloud” is the “data center” with a large number of servers and storage. Since the current semiconductor chips and other supporting hardware are still very power-consuming, the power of global data centers is equivalent to the power of 30 nuclear power plants, and what is sad is that 90% of the power is wasted because of the high efficiency. low. Google’s electricity consumption in data centers around the world has reached 300 million watts, which is more than the electricity consumption of 30,000 American households.
If the cloud computing center that now uses a lot of electricity to maintain can still provide Internet cloud services to the majority of users, when the amount of data transmission further increases exponentially, this cloud center may not be able to continue. This data transmission refers to the transmission between a large number of wireless terminals and the “cloud”. With the advent of the Internet of Things, various household appliances and a large number of sensors, including sensors embedded in wearable devices, will be connected to the Internet in the future, resulting in extremely large amounts of data. The sending and receiving of a large amount of data may cause an I/O (input and output) bottleneck between the data center and the terminal, the transmission rate will be greatly reduced, and even cause a large delay.
So someone came up with a method to add another layer between the terminal and the data center, called the network edge layer, such as adding a small server or router with storage, and storing some data that does not need to be placed in the “cloud”. This layer of direct processing and storage can greatly reduce the pressure on the “cloud”, improve efficiency, and also increase the transmission rate and reduce latency. This method is called “fog calculation”.
The name is still very imaginative. The “cloud” is floating in the sky, but you can see it but you can’t touch it, and the “fog” is right in front of your eyes, grounded, and a realistic object. The equipment used in “fog computing” is a small server or router, which is located between a large data center and end users, and can be placed in communities, factories, enterprises, homes, etc.
Effects and characteristics of “fog computing”
The name “fog computing” was originally coined by Professor Stolfo of Columbia University in New York, but his goal at the time was to use “fog” to block hackers. Later, Cisco of the United States took over this term and used it to promote products and network development strategies. IBM has been promoting the “marginalization” of the network and is preparing to gradually move many “cloud computing” tasks to “fog computing.” It is to move the data in the “data center” at the center of the network to the edge of the network.
“Fog computing” has obvious positive effects for enterprises: a large amount of internal data of enterprises does not need to be transmitted to the “cloud” and then back from the “cloud”, but directly processed through the “fog”, which can greatly improve the efficiency of the enterprise . For individuals, if the software in the mobile phone needs to be upgraded, there is no need to go to the “cloud” to upgrade, just upgrade through the “fog computing” equipment in the nearest place (such as in the community).
The main features of fog computing:
★ Very low latency. This is of great significance to the booming Internet of Things. In addition, online games, video transmission, and augmented reality also require extremely low latency.
★Large geographical distribution. This is in sharp contrast with cloud computing (data center) concentrated in a certain location. For example, if you need to send information and video to a high-speed moving car, you can set up wireless access points along the highway.
★ Large-scale sensor network with a large number of network nodes to monitor the environment. The smart grid itself is a large-scale distributed network with computing and storage resources, which can serve as a good application example of “fog computing”.
★Support high mobility. For fog computing, mobile phones and other mobile devices can communicate directly with each other, and the signal does not have to go to the cloud or even the base station to go around, so it can support high mobility.
★Support real-time interaction.
★Support diversified hardware and software equipment.
★Support cloud online analysis.
Here is an example of an “intelligent transportation system” using “fog computing”. This system uses traffic lights as network nodes that can interact with sensors. Sensors can detect the presence of pedestrians or cyclists, and measure the distance and speed of an approaching car. Through “fog computing”, these smart traffic lights can coordinate with neighboring smart traffic lights, can warn approaching cars, and even change the cycle of traffic lights to avoid traffic accidents. The data in the fog computing server of the intelligent transportation system will be transmitted to the “cloud” for global data analysis.
Another example is the wireless sensor network that has been used for many years. It is characterized by extremely low power consumption. The battery can be replaced every 5 or 6 years, and it can even be powered by solar energy or other energy sources without using batteries. Such network nodes have only very low bandwidth and low-end processors, and small-capacity memory. The sensor mainly collects environmental data such as temperature, humidity, rainfall, and light intensity. It does not need to be transmitted to the “cloud”, but only to the “fog”. This will be a typical application of “fog computing”.
“Fog Computing” and “Cloud Computing”
“Fog computing” provides geographical location information of local node distribution, and the delay of information transmission is very low, while “cloud computing” provides centralized global information. Many applications require both “fog” and “cloud”. For example, data analysis in big data technology first processes and filters the sensed data through M2M (machine-to-machine communication), and then puts it on HMI (human Computer interface) This layer is textualized and visualized, allowing users to clearly understand all analyzed data.
Here “fog computing” plays a role of a frequently used “database”, and the “cloud computing” center is a place where files are stored for a long time. Due to the extremely low latency of “fog computing”, such big data analysis can be done with mobile phones, truly “mobile data analysis”.
Under the general trend that the Internet of Things will become the next-generation Internet, some of the inherent shortcomings of “cloud computing” are becoming more and more obvious, such as the inability to support high mobility, geographic location information, and high latency. “Compute” can just solve these problems, while at the same time it can bring a lot of new applications and new services.
The nodes of this “fog computing” can be wireless routers and set-top boxes that people are using today. Before the concept of “fog computing” was proposed, in 2009, someone proposed the ideas and products of “Cloudinabox” and “Cloudlet”, which were used in peripheral mobile phones and various mobile devices. The idea of ”fog computing” is the same. For example, augmented reality is to superimpose a layer of information on the image of the real world that is actually seen. It requires real-time and complex video operations. Because it is a real-time requirement, it requires extremely low latency. Some people use Google glasses and paired them with “Xiaoyun” devices to achieve a good user experience.
Cloud computing requires a lot of bandwidth, while wireless network bandwidth is limited. The use of “fog computing” can greatly reduce the amount of bandwidth required. In principle, the transmitted data can be “bypassed”, that is, bypassing the Internet to make these data as local as possible. The most valuable data can still be transmitted through the “cloud”, but most of the data traffic can be diverted from these networks, thus greatly reducing the traffic burden on the cloud network. Intelligent transportation systems, smart grids, smart cars and even health care systems, etc., if they all process data locally and only transmit the most important data to the “cloud”, it will become very efficient and save a lot of costs.
But “fog computing” does not compete with “cloud computing”, but can be seen as an extension of the latter. In this sense, some companies use drones and hot air balloons equipped with wireless communication equipment to provide broadband network services to remote areas, which can also be regarded as a kind of “fog computing”, and this kind of “fog computing” is true. Very vivid: floating between the user and the cloud, a bit of “fog” smell.
“Fog computing” has many advantages, but when it comes to large-scale deployment of “fog computing” equipment in the future, various problems will inevitably be encountered, such as network congestion, software architecture, security, privacy, and what business model to adopt, but these problems It will definitely be resolved in the near future.
“Cloud computing” and “fog computing” have appeared in the dictionary of computer technology. Can we expect “snow calculation”, “rainbow calculation”, “hail calculation”, and even “storm calculation” next?