China Mobile Sun Shaoling: "Big cloud" faces four technical problems

China IDC Industry Alliance News in today's "2010 Mobile Internet Symposium", China Mobile Communications Institute, director of Cloud Computing Institute Sun Shaoling briefly introduced the "Big cloud" research and development and practice. In his view, telecom operators have a wealth of network and infrastructure resources, can effectively protect the reliability of cloud computing and service quality.

According to Sun Shaoling, China Mobile is quite active in cloud computing, in order to achieve the future IT infrastructure and mobile internet infrastructure, China Mobile started its research and development project in 07, built a 256-node cloud computing cluster in 08, and expanded the lab system to 1024 nodes by the end of 09, And in May this year, the official release of the Big Cloud 1.0 results.

In the specific framework, Sutra currently 1024 nodes are placed in a rack-type server, with a total capacity of more than 5,200 CPU combination, 10P memory, 2.8P hard drive, China Mobile all deployed open source system, at the same time built a number of large-scale data storage management of cloud computing systems.

Sun Shaoling also introduced some technical difficulties in the development of China Mobile's "big Cloud".

First, there is a usability problem, in the large cloud platform currently used in the technology and parallel computing systems, there are problems in the design. Next is the massive data processing aspect, the China Mobile Sutra has the limitation in the massive data processing, all data output input All is the dependence file, facing the performance relatively high file sometimes cannot satisfy the demand.

Third, distributed storage architecture and read-write optimization, divided into the current distributed architecture processing transaction business model, relatively easy, but in the relational database, there are some deficiencies. In the aspect of high scalability, it faces the problems of massive file storage expansion, database extensibility, scale of virtualization resource pool and interconnection among multiple data centers.