Cluster storage will be the best choice for high-definition monitoring

From the characteristics of modern security monitoring technology, we can see that the monitoring system needs now: First, all levels of fault tolerance, that is, when the storage device fails, no data is lost, and the storage device fails to be affected when the service is not affected (automatic switching to another The storage device continues recording; secondly, it is forward-looking (that is, scalability) that can expand capacity as the volume of the service increases, and the impact on the front-end business is minimized when the capacity is expanded; third, the capacity is allocated on demand, ie, the disk is increased. Space utilization rate; Fourth, supporting multiple applications such as video retrieval, intelligent analysis and concurrent implementation; Fifth, respond to energy conservation and environmental protection requirements and reduce carbon emissions.

The deficiency and defect of the traditional IT storage architecture HD monitoring storage system needs a series of requirements such as efficient management of massive data, efficient retrieval, high reliability, on-demand capacity allocation, and energy saving, but traditional IT storage methods (including DAS storage, NAS Storage, SAN storage seems to have incompetent reasons in these areas:

DAS: This solution uses SAS or HBA cards to connect directly to application servers and fiber arrays without providing data sharing capabilities. If multiple applications need to share the same data, it often takes a lot of time to migrate the data. As a result, there are multiple copies of the same data in the environment. Synchronization of multiple data is difficult, which wastes a lot of manpower and material resources and is difficult. Ensure data reliability.

SAN: This solution utilizes storage fiber networks to connect application servers and fiber arrays, which is more flexible and scalable than DAS. However, SANs do not have data sharing capabilities, and application servers are difficult to be redundant. Once application servers or disk arrays fail, front-end services are interrupted, and the high price of SAN systems and device compatibility among different vendors are limited. The use of high-definition monitoring systems.

NAS: NAS systems that provide NFS or CIFS protocol access can provide application servers with a unified file system image to meet the need for data sharing among multiple application servers. However, because traditional NAS often provides file-level storage space in the form of a single server, its limited IO bandwidth and scalability are difficult to meet the large number of concurrent read-write needs of high-definition monitoring systems. Moreover, the NAS has poor support for file system locks, and multiple clients cannot read and write the same file at the same time, resulting in the inability of clients to work together well.

SAN+NAS: This solution has a certain degree of SAN scalability and NAS data sharing capabilities, but there are still obvious flaws. All application data I/Os must be performed on a single NAS server. Server load balancing cannot be implemented. As a result, the I/O bandwidth of the entire system is limited to the bandwidth of the NAS server, and it is still difficult to meet the needs of high-definition monitoring system environments.

Faced with this situation, although everyone put their hopes on cloud storage, however, cloud computing is still early to be used, and the time for cloud storage applications to be deployed remains to be discussed. How can security monitoring users implement high-definition data storage for large amounts of data? ?

Combining the unique application requirements of the modern surveillance market and various defects and shortcomings of traditional storage architectures, the domestic blue-and-white whales of network security monitoring storage experts have launched a cluster storage solution in a comprehensive manner for different users' different networking methods.

The so-called cluster storage refers to aggregating the storage space of multiple storage devices into a unified named storage space. Applications can transparently access and utilize the disks on all storage devices through the access interface. China Blue Whale BWStor cluster storage realizes the consolidation of unified storage space through system file-level virtualization, greatly improving the performance of storage devices and disk utilization, enabling capacity dynamic on-demand distribution and online without affecting front-end services. Expansion. In order to obtain higher concurrent access performance.

Clustered storage is mainly embodied in improving the overall performance of parallel or partitioned I/O. In addition to realizing high concurrency performance, the BWStor clustered storage system of China Blue Whale realizes environmental protection and energy conservation. That is, according to the smart storage strategy, hard disk partitioning works in turn. When the disk does not work, it stops spinning, which reduces overall costs and extends disk life.

The traditional type of storage architecture, the average I/O read/write speed per 1TB hard disk can reach about 14-21MB/s. With 50TB as an example, the I/O concurrent read/write speed of the entire storage architecture can only reach 0.75-1.25GB. /s cannot meet the requirement of high concurrency; while the BWStor cluster storage system of China Blue Whale has broken the traditional bottleneck of I/O reading and writing speeds in hard disk units, and adopted an efficient file management mechanism. The node is the unit of calculation for reading and writing the entire system I/O. Similarly, with 50 TB, the BWStor cluster storage system of China Blue Whale requires approximately 9 nodes (9 storage units), and each node has a read/write speed of 300 MB/s or more. Therefore, concurrent read and write speeds can reach 3GB/s or more, which greatly meets the user's high concurrency requirements. After testing, the SkySky 1500 of China Blue Whale supports 190 channels of 4M high-definition code streams and read and write concurrently. The cluster storage system composed of nine SS1500s supports up to 1700 channels of 4M high-definition code streams concurrently for reading and writing.

Through analysis, it can be seen that when the software of the traditional storage architecture has a large number of client requests, the metadata operation will lose its response. The BWStor clustered storage system of China Blue Whale does not have this problem, because it adopts a separate advanced mode for recording and broadcasting, which can effectively balance the load, thus solving the problem that the server loses its response due to all clients requesting too much data at the same time. The BWstor cluster file system of China Branch Blue Whale also supports device-level fault redundancy, namely N+N redundancy scheme. When the storage device being written fails, the system will automatically switch to the next storage device, thus guaranteeing the client's request. Data is not affected, greatly improving the reliability of the data. After testing, under the same hardware conditions, the cluster system software IO bandwidth is 35% higher than traditional storage architecture software. Investment costs have been reduced by more than 45%.

Opportunities for cluster storage Currently, the domestic surveillance industry is in rapid development. The development of the security market is mainly driven by some government-led projects, such as the construction of safe cities and the construction of traffic monitoring systems. These security systems all have multiple paths. The requirement of high image quality requires the system to have the ability to process information in real time and the ability to intelligently analyze deep data mining. The problem of storing large amounts of data is inevitable.

The development trend of modern high-definition surveillance is digitalization, networking, integration, intelligence, and standardization. Advanced security technologies continue to penetrate into various industries and form distinctive applications in different industries. Which cluster storage for the security industry, this new storage method can be accepted by us? I believe that through a certain period of familiarity and understanding, the majority of users will accept.