HOME > Solution > Solutions >

Distributed storage solution based on Intel Scalable Process

TIME:2020-04-10 14:26       

      Beijing Jinpin High-end Technology Co., Ltd. was incorporated in Beijing Zhongguancun, known as China's Silicon Valley in April 2000, and is a manufacturer and solution provider of domestic servers, GPU / deep learning, and storage products. With nearly 20 years of experience in this field, the main business includes independent brand "jinpin" series servers, workstations, storage products, ODM / OEM customized services, GPU deep learning, high-performance cluster computing, distributed storage, hyper-convergence , Video retrieval analysis, emergency disaster recovery backup and other solutions.


background

      In today's society, high-definition pictures, videos, community networks and other Internet applications are in full swing, people's communication, work and lifestyle are constantly changing, everything is becoming digital, we are not subject to audio, video, pictures, E-Mail all the time , Various electronic documents, reports, bills and other information, and unstructured data grows exponentially by 60% every year, so companies have to use more disk arrays to increase storage capacity, but it seems that the growth of data is always manufacturing Trouble, cost increases, the storage system becomes more complex, and storage performance will obviously decrease when the amount of data is getting larger, and the security of data will be more challenged. How to store these unstructured data has become a problem faced by enterprise IT departments.


Solution

Jinpin Kstor cluster storage system is a large-scale universal cluster storage system designed and developed for massive unstructured data storage with 100% domestic independent intellectual property rights. It uses massive virtualization technology, distributed storage technology and cluster technology The server is virtualized into a shared hard disk with large capacity, high performance and strong scalability, which provides data read and write access for the front end.



Features:

     Distributed architecture, unified namespace, clustering of metadata and storage nodes;

     Supports modular expansion, storage nodes can be expanded by 10,000 units, metadata nodes can be expanded by 256 units, and the capacity can reach EB level;

     Support multiple storage protocols at block level, file level and object level;

     Multi-node aggregate bandwidth up to hundreds of GB / s;

     Support N + M erasure code, multi-copy data protection mode;

     Support hierarchical storage, load balancing, fault self-healing and other functions.



At present, most storage systems on the storage market have the problems of low throughput and slow retrieval under the application mode of massive small files.

The use of the second-generation Intel® Xeon® Scalable processor significantly improves the efficiency of small file read and write queries:


Typical applications


SALES:(+86) 135 5249 6467    yj@jpgd.cn

ADD:4th Street, Shangdi, Haidian District, Beijing, China

Inquiry Us Now !

Copyright (c) 2020 jpgpu.com All Rights Reserved