Customer’s comment

Director Yang of Suizhou Business Division of Zhongke:
vClusters’ fully integrated distributed storage solution has been running for three years without any failure, undertaking the storage and forwarding of the data of violations of laws and regulations and throughput at the traffic police bayonets of Guangshui City. Based on the demands of continuous power supply and uninterrupted businesses, the solution has completed phase-2 capacity expansion this year, which will greatly reduce our company's input in operation and maintenance and system construction costs. vClusters’ service attitude and technical support are very satisfactory throughout the process, and we also look forward to further win-win cooperation with vClusters in more projects.

Customer demands

"There’s an over speed violation camera 200 meters ahead with a speed limit of 120 kilometers, please slow down!" Every time passing through an intersection, you can always see a lot of cameras, and most drivers will be familiar with such a voice reminder. As the Safe City progresses, traffic system building is becoming increasingly intelligent. Amongst others, the traffic police bayonets play a very important role. They assume important tasks in vehicle behavior monitoring and also in easing traffic congestion, improving road traffic efficiency, and criminal case detection and can effectively curb the phenomenon of traffic violations and eliminate traffic hazards.

Traffic police bayonets take a large number of pictures daily, so terabytes of small picture files need to be processed every day; and what’s more, these data need to be stored for a certain period of time. Hence, hundreds of millions, billions, or even billions of millions of small files have brought great challenges to the bayonet system. However, most of the storage systems currently available on the market are designed for large files, and they lack targeted optimization for the storage and reading of massive small files, so they are not suitable for scenarios with massive small files to be stored, such as traffic police bayonets.

Pain spots of users

Several companies in the industry are working to solve the problem of massive small files. vClusters is just one of them. While studying the special scenario of traffic police bayonet, vClusters has found that the massive small files mainly have impacts as follow on the bayonet system:

● Massive small files are slow to delete!

A general small and medium-sized city takes tens of millions of images every day, and each of them has an average size of about 200KB only. However, as the traditional storage system has very low efficiency in deleting them, the storage space cannot be released in time, which often affects the writing of normal business data.

● Massive small files are slow to read, write, and retrieve!

As the traffic police bayonet system has stored a large number of small files, while read, write and retrieval operations are carried out, the metadata are frequently accessed, and the magnetic head needs to seek and change the track frequently, resulting in a very large latency. In the case of the peak period, high concurrent access may even cause system collapse.

● Massive small files are hard to share!

The important pictures collected by the bayonets need to be uploaded for provincial and municipal institutions to read at any time, query quickly, and browse without any lag. Also, different permissions shall be set up for the server and the client, for example, the server has all the permissions while the client can only read and browser data but not delete and create data.

vClusters’ advantages

vClusters has its unique solutions to the problems of massive small files. Through repeated researches and tests, the company's research team specially optimized the technologies for the massive small files and launched vCluster distributed storage system and Nxcells next-generation micro-cluster solution in the end. The vCluster distributed storage system can provide high- throughput concurrent writing of data and concurrent reading and writing of multiple channels of videos, thereby solving the problems of massive small files in retrieval and deletion. The NxCells next-generation micro-cluster solution can provide less space density and lower energy consumption than traditional storage, thus greatly reducing the operation and maintenance costs of machine rooms.

● Rapid deletion of massive small files!

With regard to the deletion of massive small files, the vCluster distributed storage system has provided a storage that can hold huge amounts of data. With each single volume capable of providing petabytes of storage space, the system can manage billions of files efficiently, and its directory can support tens of billions of files. Therefore, it can realize an extremely fast detection speed of 200,000 small files per second, which is much faster than some foreign manufacturers in the industry in the competition tests.

● Efficient read, write and retrieval of massive small files!

As claimed by vClusters in its massive small file solution, the vCluster distributed storage system itself is capable of analyzing and processing massive data. Through the small file merging technology, it can have multiple small files merged and stored in a large file to reduce disk and network overhead and realize efficient small file storage. Through the efficient massive small file retrieval technology, it can maintain an efficient retrieval speed while writing massive small files.

● Global data sharing and user rights classification!

Relying on the global name space of the file system, the vCluster distributed storage system can present a consistent view of the file system to all computing nodes and archive all important data to the cloud to realize the centralized management of all documents and collaborative office. In such a way, institutions at all levels can access archives at any time via terminals to realize the data resource sharing across departments and levels, and read and write permissions can be set for each user role to facilitate the management of personnel permissions.

● High density, low power consumption, and reduced cost!

In addition, vClusters’ NxCellsnext-generation micro-cluster solution has very low power consumption, which is in line with the national policy of developing green data centers. With the help of the innovations in product hardware design, the Nxcells’s capacity doubles that of a traditional X86 server for the same machine room space, namely for the same storage capacity, Nxcells can save 50% room space. Therefore, it can greatly reduce the costs of power, refrigeration, space, and operation and maintenance, etc.