Wuhan University: “OGON” Big Data Platform Based on SandStone
As a comprehensive and key national university directly under the administration of the Ministry of Education of the People's Republic of China, Wuhan University is also one of the "211 Project" and "985 Project" universities with full support in the construction and development from the central and local government of China, included in the first batch of the universities shaping themselves into world-class comprehensive research universities domestically and internationally and one of the “most prominent institutions of higher education in China” listed by the world renowned journal Science.
In order to store and manage its massive unstructured data generated by teaching, scientific research, management and other related IT systems in a unified manner, and in response to the call of the national “smart campus”, Wuhan University Information Center has planned to build a “big data cloud storage platform” to meet the needs of long-term data storage and management. The Center has carried out a multi-dimensional discussion on the storage demand, mainly reflected in such four aspects as follows:
1) There is a great amount of unstructured data produced by scientific research and teaching and such data is growing fast, so the storage solution must meet the needs of linear expansion;
2) The unstructured data is large in capacity and there should be a perfect solution for replacement of the old equipment, so as to avoid “manual” relocation of the massive data;
3) The storage solution should be able to unify the name space, so that the IT operation and maintenance could be more convenient after the platform is released online later;
4) In order to meet the requirements of hybrid cloud construction in the future, it is hoped that the storage protocol should adopt such a protocol compatible with public cloud storage.
After 12 months’ strict assessment, Wuhan University Information Center finally selected the object storage solution provided by the famous manufacturer - SandStone in the field of software-defined storage through multiple evaluations on the mainstream storage solutions in the market, including well-known manufacturers’ medium and high-end NAS storage arrays and many manufacturers’ object storage solution.
As the optimal solution for storage of massive unstructured data, SandStone object storage solution has adopted a decentralized and distributed technological architecture, with linear scalability available for storage capacity and performance. Moreover, the solution is competent for storage of tens of billions of files as well as massive data analysis and retrievals, and read-write delay can be controlled within milliseconds. At the same time, it can support the S3 file access protocol and users can set access policies to provide web storage services and easily realize data storage and management in thehybrid cloud scenario.
Data online analysis has been accelerated from 2 hours to less than 30 minutes.
Flexible expansion is available to realize on-demand construction and linear increase of storage resources.
Support multi-terminal access and improve access efficiency.
The storage system is decoupled from the container, so the change of the underlying storage pool has no impact on the container.
File lifecycle management is provided, files can be archived automatically and historical data can be previewed online.
The storage system is highly available. When individual nodes or hard disks fail, the business continuity can be guaranteed.