High-definition surveillance market: NVR will become an important storage form

As high-definition surveillance is increasingly used in the market, its role is becoming more and more important. Let us first analyze its characteristics.

    Analysis of NVR application characteristics

    First, NVR realizes front-end access based on IP ports. Its access capacity varies from a few channels to dozens or even hundreds of channels, and there will be various forms. However, due to the limitation of the number of analog ports, the main form of large-capacity DVR is 16-channel access. Therefore, NVR has obvious advantages in multiple capacities and large-capacity concurrent storage, and can meet various storage needs.

    Second, NVR is equipped with advanced storage reliability technologies such as RAID and ANR since its birth, which has a unique advantage in ensuring data integrity. At present, most DVRs do not have RAID function, and due to its analog access characteristics, ANR technology will not be used. ANR technology is becoming a very important reliability technology in IP surveillance storage. It realizes the perfect integration of local storage and network storage: first, it detects network status. When the network fails, it can record the transmission data before the failure, and upload the data to the management center after the network connection is restored; second, it is dual recording. When the network has problems or is unstable, both the management center and the front end will store the image data separately in the buffer for backup; finally, it is automatic data repair. When the network returns to normal, the management center and the front end will compare their respective data, check and compare the image data when the network fails, and then automatically repair the wrong or lost data in the management center.

    Third, the network openness of NVR itself also enables NVR to provide more flexible and convenient storage deployment, and realize unified management and unified retrieval of storage resources while alleviating network traffic pressure.

    Analysis of the characteristics of IP-SAN storage

    The NVR storage discussed above is mainly for small and medium-sized monitoring systems or industry terminal access monitoring systems. In addition, many large-scale, distributed network video monitoring systems currently use massive storage devices such as IP-SAN. Below we will analyze the reasons in detail from the perspective of technical trends.

    The traditional IP-SAN storage system architecture is as follows: after the network camera or video encoder encodes and compresses the collected video, it is uploaded to the central management platform (for a multi-level system, it is a sub-center management platform) in the form of an IP video packet. The central management platform or sub-center management platform then forwards the IP video packet to the recording server, and the recording server processes the IP video packet and then sends it to the IP-SAN disk array for storage. When the client or TV wall wants to access and replay related recording resources, it goes through the reverse process of IP-SAN-recording server-platform-client or TV wall. All video code streams must be forwarded through the management platform.

    As a form of network storage, IP-SAN plus platform storage solutions are more flexible in deployment location and can have a very large storage capacity compared to previous DVR storage. On the other hand, it can achieve reasonable distribution of network traffic and improve storage reliability through distributed deployment, and it can also simplify user access under unified management. Therefore, the application of IP-SAN storage in the field of monitoring is a huge progress. Now many large-scale industry monitoring systems are beginning to adopt IP-SAN in network construction or upgrades.

    However, the IP-SAN + video server + platform architecture is not a perfect solution. It still has certain defects, which are mainly reflected in the following points:

    First, there is still room for improvement in reliability. In terms of the physical device form of "recording server + management platform", some are in the form of "slave platform + master platform" and some are in the form of "server + master platform". For small systems, even one device integrates the functions of both. Regardless of the form, the front-end video must go through at least two network transmission processes before it can be delivered to the IP-SAN, and the video stored in the IP-SAN must go through at least two network transmission processes before it can reach the client or TV wall decoder. This affects the reliability of the storage system to a certain extent. In addition, IP-SAN itself does not have a file system. The function of the file system is provided by the recording server. When a network failure occurs between the two, reliability risks will be caused by the repair of the file system;

    Second, the platform performance requirements are high. Whether it is the front-end video storage process, or the client or TV wall accessing and playing back the recorded data, the video resources must pass through the recording server/slave platform, which places high requirements on the platform's media forwarding performance and is even a waste of platform performance;

    Third, the distributed deployment of IP-SAN itself is not well implemented. This architecture actually makes the video server and the platform more tightly bound, which leads to the fact that IP-SAN still cannot implement distributed deployment well;

    Fourth, load balancing is not perfect. Whether one video server/slave platform corresponds to multiple IP-SANs, or multiple video servers/slave platforms correspond to multiple IP-SANs, it is difficult to achieve perfect traffic load balancing, or it is difficult to ensure the rationality of the load balancing strategy.

    With the penetration of CDN and other streaming media related technologies and the gradual adoption of ONVIF standards, the above-mentioned "IP-SAN + video server + platform" storage architecture is being impacted by a new storage model.

    NVR-based streaming media storage architecture

    This storage architecture integrates the video server/slave platform with the disk array to form a "new" storage form. After the network camera or video encoder encodes and compresses the captured video, it directly transmits the IP video packet to the storage device, and the bit stream is not forwarded by the platform. When the client or TV wall wants to access and play back related video resources, the video information is also directly transmitted from the storage device to the client or TV wall, and the bit stream is not forwarded by the platform. The platform is mainly responsible for the interaction of signaling and management and control services in the middle.

    What problems can NVR streaming media architecture solve for HD storage?

    First, reliability is improved. The code stream only has one network transmission process from the front end to the storage device or from the storage device to the client/TV wall, which reduces the links that affect reliability. At the same time, the streaming media is directly processed in the storage device, which is equivalent to local storage, reducing the packaging process of the storage protocol;

    Second, it reduces the loss of platform resources. All code streams do not need to be forwarded through the platform. The platform is only responsible for signaling interaction and control management. One platform can manage large systems.

    Third, truly realize the distributed management of storage devices. The storage devices and the platform can truly realize the network-based connection, can be deployed in any location, and can be reasonably deployed according to the distribution of front-end points, thereby achieving a more reasonable distribution of network traffic;

    Fourth, efficient load balancing is achieved. Video processing and disk arrays are directly connected, eliminating resource waste and complex load balancing strategies, and making resource utilization more efficient.

    In fact, this "new" storage device form of "disk array + video server" is essentially NVR. NVR is a new generation of network storage device that integrates disk array, video server and streaming media services - network video recorder.

    Therefore, in the future era of high-definition monitoring, NVR can realize centralized or distributed video storage of various monitoring systems.