Shulou(Shulou.com)06/03 Report--

With the rapid development of the Internet of things and the advent of the 2.0 era of data processing, edge computing has quickly become a popular technology. IDC data show that in the future, more than 50% of the data will need to be analyzed, processed and stored at the edge of the network, and the market for edge computing will exceed trillion, making it an emerging market on a par with cloud computing.

The strategic layout of Amazon, Ali, Huawei, Microsoft and other giants will not only provide Smart Internet services nearby in the future, but also meet the key needs of the industry in the process of digital transformation, gradually becoming another new hot point in the era of the Internet of everything. However, it should be noted that whether in the security industry or industry, there is a "close relationship" with the Internet of things on the edge of computing to land, how to apply will be the core issue. However, if we want to understand the logical relationship, we need to analyze it step by step.

What is marginal computing?

In a broad sense, it provides flexible computing services with low latency, high reliability and high availability near the user data and access side. to meet the key needs of customers in real-time business, application intelligence, data processing and analysis, data security and privacy protection, large-scale edge computing applications can be flexibly configured and managed.

What is the relationship between cloud computing and edge computing?

Cloud computing and edge computing will be a symbiotic relationship. For the need for timely response, the algorithm is relatively simple to calculate, using edge computing processing; for non-computational response, the algorithm is complex calculation, carried out in cloud computing. The future will be a combination of front-end intelligence, edge and cloud.

Front-end intelligence

With the continuous mature development of AI chip and embedded sensing system, the computing power of front-end intelligent devices is increasing, and more complex visual computing functions can be completed, so that the results of detection, recognition and classification can be applied in the front end in real time. The real-time and security requirements of intelligent analysis in specific scenarios need to be directly realized on the front-end intelligent equipment; the front-end intelligent processing can also transmit high-quality structured data and analysis results to the back end as needed, reduce information loss or errors caused by packet loss and compression, and improve the accuracy of intelligent analysis.

Cdn.xitu.io/2019/1/24/1687e8d6b8bc388c?w=750&h=527&f=jpeg&s=150282 ">

The front-end intelligent carriers are generally hardware devices with certain computing capabilities, which can achieve different intelligent functions, such as face detection, face feature extraction, face recognition, human body recognition, vehicle recognition and so on. The front-end intelligent device docks up to the cloud platform or edge computing unit to push video and audio streams and structured data to the cloud. It can also manage peripheral devices and respond to live cloud messages (can output access control information and alarm according to different cloud needs).

Edge computing

Edge computing nodes mainly collect and store all kinds of heterogeneous data of intelligent front end nearby, manage and schedule intelligent computing resources nearby, and meet the needs of real-time response and real-time analysis to intelligent analysis in different situations. For complex systems with multi-level services, edge computing nodes correspond to each hierarchical application to meet the computing needs of different-level systems. The edge computing node can receive, integrate and transmit the structured data of the intelligent front end, and it can also deploy the computing power according to the needs, and apply different algorithms to intelligently analyze the data in the current classification to realize the intelligent application.

A single edge node can uniformly manage the storage resources and computing resources needed for intelligent front-end and edge computing at this level, and schedule intelligent algorithms according to demand, combined with the intelligent analysis capability of edge computing nodes. achieve all predetermined intelligent functions at this level According to the demand, multiple edge computing nodes can be combined to form an intelligent network to process data, exchange data and share calculation results in the network.

Cloud Computing

The cloud computing platform can not only manage the front-end intelligent devices and edge computing nodes, but also complete the intelligent analysis function. The cloud platform allocates computing and storage resources according to demand, and schedules intelligent analysis algorithm and big data analysis algorithm according to business needs.

Cloud computing platform performs higher-level intelligent analysis and processing of structured data from edge computing nodes, such as face comparison with more database capacity, event correlation analysis within larger time width and space span, and so on. Big data analysis algorithm intelligently analyzes and processes the structured data returned by intelligent front-end and edge computing nodes, which supports multi-dimensional big data's comprehensive event analysis, logic analysis and decision analysis. Cloud computing platform, edge computing node and intelligent front end combine with each other to achieve multi-level intelligent analysis.

The advantage of Edge Computing in Security Industry

What are the application advantages of edge computing in the security industry (Internet of things) compared with cloud computing?

In fact, there are two types of edge computing in the market, which should be analyzed from the actual scenario requirements of the Internet of things:

50 billion devices will be connected to the Internet in the next 2020, posing two major challenges:

1. How to manage these connections?

Second, how to transmit and handle these traffic?

In order to solve the above two problems, two types of edge computing platforms are needed to provide services, one to solve the connection and the other to solve the traffic. The former mainly solves the problem of protocol conversion between the Internet and the Internet of things, and generally provides cloud services like the Internet of things gateway. The latter mainly deals with the transmission of massive streaming media or data traffic, and generally uses the edge node of CDN to provide edge computing services. Therefore, edge computing is located between WAN / sensor and cloud computing in the Internet of things technology architecture, providing nearby data processing and data access, as shown in the following figure:

What is the relationship between edge computing and security industry?

At present, cameras and camera modules account for a very large proportion of Internet of things devices. Any smart home equipment and road monitoring, all scenes that need image data collection for analysis and feedback will need cameras. The core terminal equipment of the security industry is the camera. In the past, because of the high cost of storage and equipment, the security industry was generally only used in public transport, hotels, buildings, parks and other scenarios, now home security, new retail, business centers and so on have gradually become a common scene because of the reduction of costs.

How to balance the cost, performance and intelligent analysis of private and public networks has increasingly become the key to the landing of business intelligence.

On the one hand, edge computing can efficiently deal with huge face data, crowd analysis, biometrics, commodity identification and other analysis results because it can be calculated nearby, so that the original intelligent scene no longer needs to deploy expensive and bulky hardware equipment in the field, which greatly improves the landing efficiency and replication speed of the intelligent scene. On the other hand, because of the nearby storage of edge storage service, widely distributed cameras can also store huge amounts of monitoring data nearby, providing a business experience that can be analyzed at the nearest high speed.

These two advantages of edge computing make the security industry closely linked with it, and the security scenario can be implemented better and faster under the deployment of edge computing.

How does edge computing land in the security industry?

The landing of edge computing in the security industry requires two types of scenarios:

Private network

Edge storage privatization + edge computing privatization deployment is usually adopted. The advantage of this scheme is that the private network can ensure the privacy of data, the network exit can be opened, the data can be backed up to the public network, and the public network exit can also be opened when the local computing resources are insufficient. The business is downgraded to the central computing resource for computing and processing.

Internet

Unlike VPCs, there are usually several pain points in public network problems:

Link quality problem: it is mainly the delay of communication between the equipment and the computer room of the computing center, and the network link is unreliable.

Proprietary protocols and obsolete problems: because there are historical monitoring devices in the security field, it may not be possible to upgrade embedded programs directly, and devices from many manufacturers also need to support non-standard multimedia protocols.

Resource cost: if all the data of the local camera is uploaded to storage, a lot of worthless data will also occupy the transmission channel and storage space. If you can delete useless data in close processing, it will reduce a lot of waste of resources.

Transmission limitation: especially in the process of monitoring historical data migration, long-distance transmission to the central computer room will bring great time cost loss.

Referring to the figure above, in the Qiniuyun edge computing solution, Qiniuyun creatively added the edge storage function to enable the nearest local operators to serve in the region, which solved the first problem, the link quality problem, the high-speed transmission channel to solve the fourth pain point, and the second problem through streaming storage interface and edge computing containerization access. Local computing to delete worthless data completes data sorting and solves the third pain point very well.

To sum up, edge computing can greatly improve the performance and operation of the whole industry in terms of operating costs, bandwidth utilization, packet loss rate, business delay and other indicators. At the same time, a series of features provided by edge computing security solutions, such as streaming upload, double speed playback, de-SD card, edge intelligent analysis, and so on, have prompted more and more monitoring manufacturers to join the upgrade trend.

Imagination of Edge Computing landing in the Field of Security

The security field, as the largest scene of traffic transmission in the Internet of things, is the first to improve the overall performance experience driven by edge computing. The next step to achieve business intelligence, building intelligence, community intelligence landing will become the next Internet of things flashpoint.

For example, in the application environment such as Shang Chao, we can analyze the objective information such as gender distribution and age distribution of customers, and combine the dimensional information such as stay time in the unit area. and then get the decision-making suggestions on how to deploy the relevant store location and how to concentrate the logistics service forces such as catering. In the application of buildings and residential areas, the unnecessary security personnel can be reduced, the manpower can be replaced by mechanical power, the identity of the people entering and leaving can be automatically compared, and the identity of suspicious personnel can be alerted; in the application of social security, the possible dangerous situation and security hidden dangers can be predicted according to public security, anti-terrorism, community suspicious personnel and other information combined with time and frequency information, so as to organize public security forces to carry out social management more pertinently.

The practice of edge computing in the field of security has fundamentally broken the barriers to the landing of the original "intelligent" applications, and realized the application ideas that were originally limited by computing power, transmission environment, storage environment and many other problems.

As a leading enterprise cloud computing service provider with visual intelligence and data intelligence as the core, Qiniuyun is acutely aware of the potential problems and future development trends of the cloud computing industry, and launched edge storage and edge computing products. Combined with its own in video security, machine vision and other areas of deep practice with high-quality customers, launched surveillance video edge storage solutions and video edge analysis solutions. In the future, Qiniuyun will continue to consolidate the basic capabilities of cloud computing for customers and jointly expand new businesses in the era of artificial intelligence and the Internet of things.

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.