To crack the three major security video pain points, watch Huawei's 4 major hard-core video cloud technologies

Abstract: This article comes from a speech shared by Qi Yankun, Product Manager of Huawei Cloud Video Access Service. Combined with practical experience in video access services, we discussed the trend, pain points and Huawei Cloud video access services. Exploration and practice.

This article is shared from the HUAWEI cloud community " security video cloud trend, pain points and HUAWEI cloud video access service exploration and practice ", author: audio and video butler.

As the proportion of video resources and content in the Internet continues to increase, the corresponding contradiction between the level of video resource utilization and the ever-increasing business and product requirements will become more prominent. This article comes from a speech given by Qi Yankun, Huawei Cloud Video Access Service Product Manager, in the online sharing process. Combined with his practical experience in video access services, this article discusses the trend and pain points of video access to the cloud, as well as the exploration and practice of Huawei Cloud video access services. .

The first part is the development trend of security video, including technology trends, national policies, etc.; the second part is the pain points we found in the process of practice; the third part is sharing the application practice of going to the cloud; the fourth part is focusing on the scene and some practical processes The case in the fifth part; there is a small video to demonstrate the operation process of going to the cloud; the sixth part is communication and interaction.

1. The development trend of security video

Cloudification and AI ushered in a new era in the security industry

Security is closely related to everyone's life. There are camera-related systems in the communities where you live and the office buildings where you work. These are the most basic security. From an industrial point of view, security has gone through about half a century, and the industry has been developing for a long time. From the very beginning of analog signals, the problem of "visible" or "with/without" was solved, and the Chinese myth of Journey to the West was realized. "Clair eyes, ears with wind" in the book. After that, from the perspective of the overall technology trend and business development, it still cannot meet the demands of customers. For example, if there are multiple subsidiaries of a company located in different regions, the boss wants to remotely view the subsidiaries in different regions. In the Internet age. The combination of the Internet era and the current mobile Internet has greatly improved the security experience, but this also brings a big problem-bandwidth. To transmit video to the cloud, it is necessary to solve the bandwidth problem. This is a key technology.

Subsequently, people entered the era of intelligence, and artificial intelligence is also closely related to our lives. For example, parking lots do not need to be charged manually, but only need to recognize the license plate.

Why do cloudification and AI usher in a new era? First of all, AI reduces the demand for manpower and improves efficiency. If washing machines enter thousands of households and free people's hands, then AI can free people's eyes from the field of security. The other is the cloudification level. It is certainly possible not to cloudification, but the later construction of resources includes the increasing amount of video, from standard definition to high definition, from 2D to 3D and then to interaction. This area has an increasing demand for network and computing power. The more, so "going to the cloud" is a good trend.

Just mentioned that we have entered the era of intelligence in the security field, so what role will video play in the era of intelligence?

The intelligent world is divided into three layers: perception, interconnection, and intelligence of all things. From the perspective of the perception layer, it is mainly data collection. We obtained from the data analyzed by some analysis institutions that in the digital transformation of the industry, 65% of the data is video data, which is enough to see that video is a very important part of the digital transformation of the entire industry. Another data is that in our smart world, 49% of the applications are based on video. It can be seen that video is the core of the perception of everything and is the "clairvoyance" of the smart age.

After going to the cloud, it brings several benefits: sharing, openness, intelligence, and unified management. If you don't go to the cloud, the aforementioned resources are a problem. Another problem is that a lot of hardware stacking together will cause data and applications, and software and hardware cannot be decoupled. For intelligent evolution, as the amount of video becomes larger and larger, there are more and more demands on the GPU, which is difficult to handle locally. If they are all located locally, there are companies with many subsidiaries whose equipment is distributed in different computer rooms, and it is difficult to achieve unified management of equipment. So from now on, going to the cloud can well realize sharing, openness, intelligence and management.

From the national policy level, whether it is the Ministry of Public Security, the Ministry of Transportation, the postal industry, or dangerous goods manufacturers, the country has a clear letter, hoping that companies can achieve video networking and video sharing. After interpreting the policy, we have concluded the following four levels: The first is sharing across the entire network. Only after sharing can everyone's data be unified, do AI analysis, and link and share with other devices; the second is Full-area coverage, the main emphasis is on the reasonable deployment of videos in some areas; the third is full-time availability. Recently, news about problems with some video sites has frequently made headlines and hot searches. This means that full-time availability is not available. If they can upload Cloud has a good performance in the area of availability; the fourth is full controllability, first of all, privacy. The most important thing is that data cannot be stolen. In addition, if you want to see data, you need to obtain your authorization.

Therefore, from the perspective of national policy interpretation, the whole network sharing, global coverage, full-time availability, and full-process control are important information.

2. Analysis of pain points in the cloud on security videos

We have also found some problems in the projects we participated in recently, especially the cameras purchased by some companies come from many manufacturers, and will ask us how to connect to the cloud, because each camera has its own agreement.

In order to better serve their products, many well-known camera providers will provide corresponding video access platforms. This will cause a problem. If from the customer's point of view, multiple camera manufacturers have multiple systems, from the enterprise's point of view From the perspective of this, it is relatively fragmented, which poses certain obstacles to the sharing of data. The second point is the agreement. The current video protocols include RTMP, RTSP, Onvif, RTP, GB/T28181, etc., which will cause trouble to corporate customers. The third point is that customers will report that the video freezes, the screen is blurred, the time delay is large, and the flow is often interrupted. These are all problems encountered in actual projects. The fourth point is cost. It can be viewed from two levels. First, if it is deployed locally, the demand for computing power is relatively high at this time; the second is going to the cloud, and building BGP based on the cloud is more costly.

Regarding the above challenges, I would like to share some experience with you.

First look at the protocol. The figure on the left shows that from the network layer (IP) to the transport layer (TCP, UDP), there will be many corresponding protocols for audio and video transmission. You can see from the perspective of audio and video transmission. There are at least three levels. The first point is signaling control, including H.323, which is a more common protocol used in the traditional telecommunications industry, and the SIP protocol is more commonly used at the Internet level, for the entire audio and video The transmission effect is better. There will be protocols such as RSVP and RTCP, and the other is the transport layer protocol RTP, which is a relatively low-level protocol.

There are many upper-layer protocols, such as HTTP, RTMP, HLS, GB/T 28181, etc. What do these protocols do? First of all, GB/T 28181 is an agreement formulated by the Ministry of Public Security of China (which will be highlighted below). Onvif and PSIA are also based on the security system, and are more protocols used in overseas markets.

Let's talk about a few more typical protocols. RTSP has a better delay, but it is more complicated and does not support the Web well. From these characteristics, it can be seen that it is a more suitable protocol for security. RTMP divides the data into many small blocks for transmission during the data stream transmission process, which will cause a lot of delays, about <3s, but it supports the Web better, and you can view it on the Web through a plug-in, which is more suitable Live scene. From a cross-platform perspective, HLS runs better, but the latency is relatively large.

In summary, it can be seen that different protocols have different positioning. Then, in the security system or in the business development, how to choose the appropriate protocol to meet the business development requirements, then there is the following GB/T 28181 agreement.

This agreement itself was formulated by the Ministry of Public Security of China. There are currently two versions GB/T 28181-2016 and GB/T 28181-2011. The main purpose of this agreement is to solve the problem of docking between camera systems of different manufacturers and standardize several An important scene. From this perspective, in order to solve the problem of docking of different manufacturers, the GB/T 28181 protocol is a very suitable protocol. GB/T 28181 actually defines the SIP, SDP, and RTSP just mentioned, which can be regarded as a protocol specification. The agreement mainly defines 12 scenarios. The device can register and cancel with the server, can real-time audio and video on-demand or historical video playback, device control (camera rotation direction), and have the ability to locate and notify changes in device status.
With the GB/T 28181 protocol, from the perspective of Huawei Cloud, we have implemented this protocol first. All mainstream cameras in the current project can be accessed based on GB/T 28181, which can solve the problem of multi-vendor, multi-vendor The issue of agreement. If different business scenarios have special requirements, we can also support standard protocols such as RTMP and HTTP-FLV.

GB/T 28181 is also very simple from the point of view of devices on the cloud. You only need to do some simple configuration, and you can connect to the cloud without code development. After the video is connected to the cloud through the protocol, we also provide some capabilities on the cloud, including device status management (device online and offline), video stream management (video stream activation, deactivation), automatic invitation for abnormal video streams, Combine with other cloud services to provide cloud storage, watermarking, screenshots, transcoding and other capabilities. The most important thing is AI blessing (will be introduced later).

From a video point of view, experience is a very important thing. If the video playback is stuck and the screen is unacceptable, from experience, first eliminate the end-side problem to ensure the smooth playback of the local video. After the local problem is eliminated, focus on the network reason. It seems that from the video access, the Different networks can be used for entry. The dedicated line network is relatively stable for video access, and the effect of household broadband is poor. From the perspective of the cloud, bandwidth reservation is very sufficient.

Then there is cut-off. There are many reasons, maybe because the system itself is abnormal. HUAWEI CLOUD’s solution is to use the automatic solicitation capability. The system automatically recognizes which streams are interrupted, and then automatically pulls up the streams. The customer experience will be greatly improved. Some interruptions are caused by customers who need to actively interrupt services based on business characteristics. The solution at this time is that the cloud system automatically recognizes and informs the customer through active subscription, and the specific follow-up actions are handled by the upper-level application.

Finally, the end-side configuration is complicated. In fact, the GB/T 28181 configuration is relatively simple. The camera itself is also a small system that requires login and some configuration, which may be more complicated for construction personnel. The solution is to scan code registration, and manage device information based on QR code scan code on the end side and complete cloud registration, which will intuitively reduce the configuration complexity. If the construction personnel have a better understanding of GB/T 28181, they can configure it directly on the equipment side.

Costs are of great concern to companies. The cost of video on the cloud is roughly divided into two parts.

The first part is the enterprise bandwidth cost of enterprise customers themselves, and they need to discuss with the operators. The method to reduce the cost on the access side is mainly to reduce the transmission low bit rate. The second part is the cost of cloud bandwidth. We need to talk about the cost of cloud bandwidth. Here we have to mention the native media network of Huawei Cloud Cloud. We run a lot of services on the native cloud media network, including live broadcast, on-demand, and real-time audio and video. Big Cloud also has edge processing, which is a media network composed of Big Cloud and the edge. Based on such a media network, we can reuse network bandwidth resources well.

From the perspective of cloud video access, it is divided into several parts. Based on unified cloud scheduling, signaling messages are processed by the cloud first. This is the signaling scheduling process. If the access is legal, the camera or NVR will obtain the access point from the cloud. This access point may be an address of a different edge, and it is scheduled based on a nearby strategy or a reliable signaling strategy. In this case, the cameras that were directly dispatched to the big cloud using BGP access will go to the edge first, and then the edge will pass through the dedicated line to send the video back to the big cloud for corresponding EI analysis or storage. This is the overall architecture. From a reliability perspective, if the edge access point fails, the video will be dispatched to other convenient nodes that are providing reliable services. If an extreme situation occurs and all edge nodes fail, then we will schedule the video directly to the big cloud. Therefore, from the perspective of the entire architecture, it is possible to guarantee the reliability of access and the lowest cost.

3. Huawei's application practice in video access to the cloud

I just analyzed the three problems we encountered in the project delivery process. So what can be done after the video is connected to the cloud? On HUAWEI CLOUD, especially the services that are combined with video include live broadcast acceleration. If you want to allow many people to watch the video at the same time after the video is connected, you can combine it with the live broadcast acceleration service. Huawei's live broadcast effect reaches a success rate of >99.7% and a stutter rate. <2.5%, playback delay <3s. In the cloud, Huawei Cloud’s storage capacity promises 99.995% of service availability, which has a good competitive advantage. There is also a combination point. If you want to do some on-demand transcoding after the video is uploaded to the cloud, or there is high-definition low-code conversion, we also have a supporting video service. The other is the combination of intelligent analysis services, which I will focus on next.

Here I have listed some intelligent algorithms related to daily life, divided into seven categories: face recognition, office building gates use this service; intrusion detection, community use is more, if an intrusion is found, it can identify and alert; personnel statistics Urban management scenarios can be applied to whether people gather or not; safety detection is applied to fire and smoke alarms; vehicle detection, detection of parking violations, traffic statistics, etc.; human body recognition, human body deployment control, abnormal behavior, etc. are usually used in urban management.

You can combine different business scenarios to support Huawei Cloud services.

In response to the above content, I will make a summary based on the three pain points we encountered in the early stage: multi-manufacturer multi-protocol, experience, and cost.

The first is the blessing of cloud capabilities, the combination of cloud, AI, storage, and codec.

The second is full real-time. The multi-edge media network can guarantee business continuity. If there is an abnormality, it can be automatically switched and provide automatic invitation for interruption.

The third is wide coverage and compatibility with multiple manufacturers. In some projects, mainstream cameras and NVR equipment such as Haikang Dahua and Huawei have already been docked.

The fourth is low cost and high cost performance. We can achieve low cost in combination with the media service network. The fifth is all scenarios. From the perspective of support scenarios, we have access to all scenarios, including security, transportation, parks, and supermarkets.

Let's discuss the security scenarios below.

4. Introduction of scenes and cases

From the perspective of video access services, the main scenarios include smart stores, home security, smart parks, smart environmental protection, and transportation networking.
Smart stores are generally remote patrols and customer flow analysis; home security, it seems that it will become a trend at present, to install cameras at home to observe the situation at home; smart parks, which are widely used, park visitor management, license plate recognition, etc.; smart environmental protection is also increasing The more important the scene, the follow-up will focus on the case of solid waste monitoring; traffic networking is also an important scene.
Next, introduce some cases that have been delivered.

Bright kitchen and stove are closely related to everyone's life. Parents are now very concerned about students’ meals in school. This is a big demand. Secondly, take-out is becoming more and more popular. Everyone hopes to see the food processing process. If these are arranged for human supervision, it will cause a lot of labor costs. It is impossible to reach such a large scale, so the video can be stored in the cloud through AI analysis and big data analysis, which can automatically identify abnormal and violation scenarios and present them to relevant supervision and management departments.

There are 9 main intelligent recognition scenes in the bright kitchen scene: not wearing a mask, not wearing a chef hat, playing with a mobile phone, smoking monitoring, topless, pests, camera obscuration, trash can not covered, and gloves not worn. The combination of video, AI and big data can greatly improve the efficiency of supervision.

Wisdom venues are related to children. There are more and more training venues now. Parents will record videos during the accompanying process. In this process, there is a large amount of recorded videos and it is troublesome to look through. Collecting venue camera videos, and then get permission. Enter children’s portraits below, search for relevant moments based on the portraits, and then edit and produce relevant videos to send to parents. Parents receive more exciting pictures that are easy to find and store.

Once the solid materials produced by chemical companies are stored and transported, improper disposal can cause major problems. The most intuitive Yancheng explosion was caused by improper disposal, and local video was also lost. From this point of view, the problem of video loss can be eliminated by video uploading to the cloud. This business scenario not only has video applications, but also other IoT devices that jointly sense and complete the full life cycle management of solid waste. In this way, companies can be well supervised and help companies standardize production.

5. Summary

A brief summary of the content shared today.

• The first is that security has entered a new era and entered the era of intelligence.
• The second is the benefits of video on the cloud: sharing, development, intelligence, and unified management.
• The third is that the capabilities of Huawei Cloud Video Access Service itself include support for multiple protocols such as national standards, RTMP, HTTP-FLV, etc., allowing easy access to the cloud.
• The fourth is that it can be integrated with cloud services after going to the cloud, especially algorithms, storage, live broadcast services, and on-demand services.

Click to follow and learn about Huawei Cloud's fresh technology for the first time~