.png)
Published: 2019
Author: D.Y.
Operates on the edge Edge Computing Part 1
Full analysis of edge computing technology
Edge computing is a computing architecture that moves computing locations from the center of the past to edge nodes in a distributed manner. Edge nodes are devices that are closer to users. The edge computing applications that most experts now recognize are not a direct replacement. Cloud computing is done in cooperation between edge computing and cloud computing. The edge is responsible for preliminary data processing and determines whether it needs to respond directly or further handed over to the cloud. The cloud is responsible for data storage and higher performance and low timeliness. Computing, such as the improvement of machine learning. In the era of the Internet of Things, we have more devices that need to respond in a timely manner, and a larger amount of information needs to be processed. Therefore, edge computing has emerged to achieve low latency, reduce transmission costs, and reduce cloud load. Current practical applications such as Amazon Continue to develop the chip business, hoping to transfer some of the functions of its own smart speaker Echo to terminal equipment in the future.
Cloud computing
Edge computing
Cloud computing vs. edge computing
The cloud has good storage and computing functions, but the transmission rate is relatively inadequate. Since data must be transmitted from the terminal device to the cloud data center after multiple layers of transmission, it is more time-consuming than edge node computing. Second, the Internet of Things devices will increase significantly in the future. According to IBM's statistics, there were approximately 15 billion Internet of Things devices worldwide in 2018, which is estimated to double to 55 billion by 2022, and the amount of data collected will also increase dramatically. If you still rely only on cloud computing, it will cost a lot of data to meet the scale of the Internet of Things, and the cloud will also be overloaded. Edge computing therefore has the advantages of reducing latency and sharing the huge amount of data. However, there are still some problems that need to be overcome in edge computing. Just like IoT devices, information security will be a big problem. Compared with the past cloud, information security can be uniformly improved and maintained by the data center, but when there are more computing nodes At this time, there will be more opportunities to take advantage of the emptiness, which is also the part that larger manufacturers need to actively overcome.
Edge node introduction
There is no unified place for the installation of edge computing. There are two main types:
1. Edge Server:
Using Blade Sever or industrial computers as edge computing node devices, and installing these devices close to the terminal devices, this approach is currently praised by many companies, because this method has several advantages including Integrate the data of various devices, cross-analyze, and easy to expand. Some people even propose to install computing equipment on mobile telecommunications base stations to ensure the popularity of edge computing. These edge computing devices installed on mobile telecommunications base stations can be used for real-time judgment and analysis when they are not driving on the road.
2. Terminal equipment:
Some people have also proposed that complex AI calculations can be performed directly in the terminal IoT device, which may be self-driving or various machines, but this method is not good for expansion and data integration, but it can save installation costs.
Application field
In the future, edge computing will be used in various IoT spaces, and it will also be a major technology that promotes the popularization of IoT, especially in environments that require timely response or have a large amount of data to be processed. The following is a brief list of the main application areas :
1. Smart factory:
It is the earliest application field of edge computing. Through edge computing, it can quickly determine whether various semi-finished products meet the standards in each step, and effectively supervise the production process, while predicting production quality and speed.
2. Smart city:
There will be a large amount of data in the future cities, so edge computing will be required for preliminary and timely processing, such as the use of edge computing to improve city safety, camera real-time image recognition, and the implementation of smart buildings.
3. Self-driving:
Autonomous driving has become extremely important because of the requirements of timely response, and it is also the field that most experts think that they have to use when they mention edge computing. In addition, in order to judge the surrounding road conditions, self-driving cars will install a large number of sensors around the car to collect a large number of surrounding environmental data images. The processing of huge data makes edge computing extremely important, which shows that edge computing is very important for self-driving cars. The promotion of will become indispensable.
4. Logistics Center:
The daily volume of goods handled by the logistics center is increasing. In the past few years, automation has been continuously promoted to reduce labor costs. In the future, edge computing can be used in conveyor belt management, goods distribution, and sorting machines to speed up the picking speed. .
Market size
Edge computing is a method to solve data transmission and to overcome the huge data processing needs of the Internet of Everything in the future. IBM estimates that there will be 15 billion Internet of Things devices in 2019, which will increase by more than three years to 2022. Triple to 55 billion. The market size of edge computing is estimated to be 1.272 billion U.S. dollars in 2018 and 6.959 billion U.S. dollars in 2024.