Analysis of mainstream technologies in current storage applications in the security industry

As security technology develops day and night, which storage technology will be the most popular today or for a long time to come? What are the current mainstream technologies in the market? This article will answer your questions.

Security storage requirements

First, security storage requires massive storage. Monitoring data is written 24/7, and the storage time ranges from 7 days, 15 days, or even 1 year. The amount of data increases linearly with time. For a safe city project, the total number of monitoring channels is usually from several thousand to tens of thousands, and projects with thousands of TB are common, which is unimaginable for the traditional storage industry.

Secondly, performance requirements. Video surveillance mainly involves writing video streams, and performance is measured by how many streams the storage can support (usually 384Kbps-2Mbps). In the case of multiple concurrent writes, bandwidth, data capacity, cache, etc. are greatly affected, and storage pressure is high. At this time, storage needs to be optimized specifically for video performance.

Third, the price of security storage is sensitive. The massive storage in the security monitoring industry has a large total capacity, which causes the total price cost to rise. Conversely, the requirement for unit capacity cost (price per TB) is very high.

Fifth, the requirement of centralized management. In security monitoring applications, since a large number of storage devices are required, the massive data in the storage devices must be effectively managed to provide convenient, reliable and transparent support for end users.

Finally, networking requirements. TCP/IP network is the network foundation for the development of security monitoring technology towards networking. Storage technology based on TCP/IP network will play an irreplaceable role in the networking process of security monitoring technology.

Based on the understanding of the above applications, storage gives us a conclusion worth learning from: security storage in the monitoring field may develop in two major areas. First, for small and medium-sized applications with not too many nodes, NVR is the main development direction in the future; secondly, for large-scale monitoring systems, video management platform + SAN storage (mainly ISCSI storage) is still the only choice for large-scale monitoring systems.

Security storage technology

1. DVR storage. DVR storage is currently the most common storage mode. The codec device is directly connected to the hard disk, and can currently carry up to 8 hard disks. However, due to the performance limitations of the codec device, the hard disk sequential write mode is generally adopted, and RAID redundancy technology is not applied to achieve data protection. As the capacity of hard disks continues to increase, the probability of key data loss due to single-chip hard disk failure is also increasing simultaneously, and the limitations of DVR performance also affect the sharing and analysis of image data. The characteristics of this method are: cheap, easy to use, and can be operated through a remote control and keyboard; the DVR method is suitable for small-scale, distributed deployment. Domestic DVR products such as Hikvision and Dahua are already very mature.

2. Encoder direct storage. Codec external storage devices are connected through the codec's external storage interface, and are mainly expanded using storage protocols such as SATA, USB, iSCSI, and NAS. This method can achieve further expansion of the codec capacity, which is suitable for the deployment of small and medium-sized security storage. The reliability of surveillance video data is guaranteed to a certain extent through RAID technology.

The direct connection mode adopted by SATA/USB mode cannot be shared and has low expansion capability, so its application is gradually being eliminated; IP network (iSCSI and NAS) mode has better expansion and sharing capabilities. For example, the series of network storage products can directly write video data with encoding devices of multiple mainstream manufacturers, reducing the server transfer link, improving performance and saving users' investment. However, this method needs to rely on streaming media servers for data forwarding and retrieval, which is prone to bottlenecks in the streaming media forwarding link. At present, direct writing usually adopts NAS storage mode. Due to the file protocol of NAS itself and other reasons, it is not efficient when writing data concurrently on multiple nodes.

3. Centralized storage. The server connects to the front-end codec, downloads data through the streaming protocol, and then stores it on the storage device. The server and storage device can be connected through SAS, iSCSI, NAS, and FC protocols. Centralized storage is suitable for the deployment of large and medium-sized platforms. Among the centralized storage methods, IP connection mode (iSCSI) and FC connection mode have good scalability and manageability, and are currently one of the most widely used methods. In terms of actual deployment and effects, FC storage is more suitable in monitoring environments with more nodes due to its powerful performance and data processing capabilities; while IP storage has higher cost-effectiveness in small and medium-sized applications due to performance and scalability limitations.

4. NVR storage. In the video surveillance system, NVR is an ideal upgrade product for analog video recorders and hard disk video recorders. It is a stand-alone independent operating device based on the original DVR that does not require the Windows operating system and computer. Since NVR adopts highly integrated chip technology and has advanced digital recording, storage and playback functions, it does not require the replacement and storage of videotapes, and does not require computer cooperation and daily maintenance. Therefore, it can achieve high resolution (up to Full-D1 resolution), high-quality real-time monitoring, and is simple and easy to use. Simply put, the security storage of the NVR system digitizes traditional video, audio and control signals, and transmits them on the network in the form of IP packets through the network interface on the NVR device, thus realizing the networking of the system based on the DVR.

In the application, although the NVR system has the advantages of fast computer processing ability, digital information anti-interference ability, easy and fast query of records, clear video images and single machine display of multiple images, etc., in essence, NVR not only fails to solve the defects of analog transmission in DVR system, but also fails to solve more management problems caused by network transmission of video stream. In fact, each NVR forms an independent monitoring center, which makes the realization of network-wide monitoring more complicated. There is still room for improvement in aspects such as remote control, multi-level scalability and networking capabilities.

Video application storage features

In some small and medium-sized video surveillance systems, the performance bottleneck of the system often occurs at the video server rather than the storage device. The bandwidth provided by the storage device is often two or three times the actual total bandwidth of the video surveillance, and the IOPS is about 3-5 times the actual demand. The storage device controller is in a half-load working state for a long time, and a large amount of high-performance resources are wasted.

To ensure high availability, efficiency, stability and security, the controller part of the storage device is generally specially designed and uses a dedicated processor and cache. Its technical parameters and performance are higher than those of ordinary PC servers. In order to effectively utilize the surplus storage resources, special application software can be embedded in the controller part of the storage device. The storage device not only provides a data storage server for the system, but also provides a certain software application server. This type of storage device is called application storage.

Application storage architecture For video surveillance network storage systems, application storage devices are a new type of high-performance storage device with media management and dedicated software application functions.

Compared with ordinary storage, application storage not only has basic data storage and security protection features, but also has network management, media asset management, video transcoding, video on demand and storage sharing management.

Application storage refers to running certain application software services on storage by leveraging the high reliability and stability of storage devices when the performance provided by the storage device itself far exceeds the system requirements. Compared with ordinary storage devices, application storage not only adds many software applications and service functions, but also will change the architectural design of many application systems.

Storage devices are generally designed based on the principles of modularity, redundancy, and hot-swappable support, eliminating internal single points of failure. Their security and stability far exceed those of ordinary PC servers, application software, and server internal storage used in the system, and can provide the system with higher stability and system security.

Efficiency: The core of storage devices, namely the controller, generally adopts a specially designed bus structure, a dedicated processing chip and cache, and optimizes data verification and data transmission. In the case of long-term operation, application storage can not only ensure the security and stability of software and server functions, but also ensure higher operation and computing efficiency than ordinary PC servers, and provide higher service quality for the system.

Simple system architecture: Due to the concentration of many software applications and service functions in the application storage, a large number of PC servers and network connection devices and network interface cards connected to them can be saved when designing the system structure, greatly simplifying the system structure. The simplification of the system structure can optimize the system workflow and improve the system's operating efficiency. At the same time, by reducing the types and number of equipment required by the system, the system construction cost can also be greatly reduced.

Saving bandwidth: If application storage is used, video transcoding and other software or services are run on the application storage device. During the entire process of video file transcoding, data is only transmitted between the controller and disk inside the storage device. No external server is required, and no network equipment is required for transmission and exchange, so it does not need to occupy the bandwidth of the network system. External servers and workstations can obtain more network data for other operations.