Dawning Shaozong: High-performance computing and large data fusion

Recently, Shuguang Vice President Shaozong in the media interview said: "High-performance computing perfectly fit the large data in the computational power, high-performance storage and other aspects of the demand, while large data for High-performance computing provides a new commercial market, help to promote high-performance computing to expand the new market blue sea, The close convergence of the two in the industrial ecological chain can promote the in-depth reform and development of information resource Organization mode. ”

High-performance Computational solution http://www.aliyun.com/zixun/aggregation/8213.html "> Challenges for large data applications

Although large data technology and large data applications have been rapidly expanded, but there are still many problems, which is more prominent is the "4V" Challenge (capacity, type, speed, value). At present, the capacity of large data can reach the PB level, the storage and processing of these large capacity data will make a great burden on the user's computing system, and the systems built on the basis of general computing ability often cannot meet the needs of large data, and can not gain the value for the users to achieve the advantage of leading

"High performance computing can play an important role in coping with big data challenges," says Shaozong. From the point of view of data storage and processing, high performance computing system can not only reduce the cost of mass storage greatly, but also process data, transmit data and recover data much faster than normal computing system. It may take 48 hours to recover a single hard drive in the past

However, the storage products combined with the dawning High performance computing technology can control the time in more than 10 minutes, significantly reduce the pressure of the users in the large-scale cluster storage system and improve the system efficiency. ”

In addition, high performance computing can provide a solution to the problem of data assimilation in large data applications. In the large data acquisition system, it is possible that the data collected by each acquisition device is different in format and type, and it will consume a lot of computational resources to unify the data. and high performance computing has made significant progress in data assimilation, its assimilation speed and accuracy has been maintained at a very high level, its development results "grafting" to large data will produce great value, and promote the implementation of data organization and management, infrastructure, decision support and Automation interface and analysis.

High-performance computing and large data will be deeply integrated to drive practical applications

For any technical concept, application is the ultimate value of its final foothold. At present, large data is in the transition from concept to application, large data application has many obstacles in industry maturity, user acceptance and system usability. How to deploy large data applications? How does the data stored in large data produce value? How to apply the conclusion of large data analysis to the decision of enterprise operation? Due to the lack of large-scale application of mature cases and industrial clusters, financial users and other industry users have been slow to deploy large data determination.

Shaozong pointed out: "Compared to large data, high-performance computing in many industry users have been deployed and applied earlier, its deployment in the software and hardware integration platform, data storage and other aspects for large data applications lay a good foundation." In particular, high performance computing and large data are interlinked in logic, and have complementary advantages in platform expansion and application landing, which satisfies the sufficient conditions for high performance computing and large data fusion. ”

In order to promote high-performance computing and the integration of large data development, improve the large data application environment, at present, the Dawn has released the "Platform integration, intelligent application," the industry's big Data strategy, and software and hardware integration platform as a support. Among them, the hardware xdata large data All-in-one machine uses the mature high-performance computer and the high performance storage system as the foundation, in the software has integrated the high performance operating system and the Hadoop software. Comprehensive collection of solutions in the light of high performance computing, high-performance storage, such as resource advantages, for industry enterprises to build competitive large data solutions and services, to achieve platform capacity and application of flexible expansion and growth, to provide real unified data analysis capabilities.