Customer profile

Beijing Zhongke Ruixin Technology Group Co., Ltd. was jointly invested and established by the Institute of computing technology of the Chinese Academy of Sciences and the core team of the high flux center of the Institute of computing in November 2014. Zhongke Ruixin has developed a series of innovative technology points for more than 10 years, forming a high-throughput computing technology system. The goal of the traditional high-performance computing model is to "calculate fast", while the typical load of high-throughput computing is the application with high concurrency and high requirements for real-time performance. Its goal is to "calculate more", that is, to increase the number of concurrent tasks processed per unit time. Based on high-throughput computing technology, Ruixin company has laid out products covering multiple levels, such as cloud, network and edge, from the aspects of system hardware, system software, application platform, application algorithm, etc., to provide customers with new computing services. With the rise of industrial Internet in the era of smart connect for all, the computing demand has changed from high-performance to high-throughput, and the high-throughput computing industry is entering a new stage of rapid and vigorous development.




In the process of reading and writing access to a large amount of unstructured data, the following barriers are encountered

  • 100+
    Number of cluster storage nodes
  • 12PB
    Total data scale
  • 3X
    Performance improvement
The complexity of operation and maintenance is high, which requires a large number of personnel to operate
In the scenario of massive small files, the storage performance decreases and the access time increases
The cluster contains both small files and TB level large files. It is difficult for the storage cluster to meet the performance requirements of both sequential reading and writing of small files and random reading and writing of large files
It needs to support pure domestic arm storage server and support access of heterogeneous clients


  • The storage capacity is more than 10PB, and 100+ domestic arm storage servers are deployed
  • File read / write access performance has been improved by 3 times
  • Highly available data redundancy and automatic recovery, reducing operation and maintenance costs
  • The expansion is convenient to meet the needs of the rapid growth of enterprise data volume