Avoiding Big Data risks: It's important not to blindly follow the target

The second "Big Data World Forum" (BDWF 2012), held at the Landmark Hotel in Beijing July 13, shows that manufacturers of database, chip, storage, and data analysis and management in the industry are actively proposing their own large data solutions, while large data technologies in finance, securities, telecommunications, The rigid demand of Internet industry is obvious, but the practical application is still in the exploration. Enterprises and providers can not be ignored is to dig bokhary data at the same time, how to prevent the risks still exist.

Now we should deal with large data, which can be highly reliable and highly scalable infrastructure, and high-performance analysis system, however, talk about the risk of large data, talk about data mining, it's the effect of how good? In fact, it needs to be validated. Avoid the risk of large data, that is, can not blindly follow suit, especially to clearly implement the goal of large data, to have practical planning, in addition to good enough data quality.

The "four V" of large data--value

Large data size, fast processing speed and a wide variety of data (volume, velocity, variability)--this is a large data well-known 3V features. However, now the experts are clear to the big data everyone on the fourth V--value, that is, the enterprise to achieve the value of large data.

Due to the existence of 3V, the implementation of the fourth V is bound to require new technology, which is also the reason for the popularity of large data concepts. To gain value, we need to build a more flexible and strong data management system, not only the database, data analysis technology to improve, the underlying architecture needs to change in due course.

Microsoft Asia Pacific Research and Development Group chief technology Officer Sun Boke that the large data end-to-end lifecycle can be divided into three steps: Managing data, acquiring storage, and protecting data. Large data requires not only management of data, but also data-rich data, whether relational, non-relational, streamlined, and ultimately insights from the data.

SAS Software Research and Development (Beijing) Co., Ltd. general manager Liu said, large data age, the requirements of storage equipment and storage mode change, storage data is to query, but only query, not analysis is a great waste. The traditional analysis ability can not deal with large data in time, that is to say, high performance analysis is the key in large data age. Liu puts forward the analysis of three ways of allocating data to different machines, analyzing the internal analysis of the analysis process and putting it into memory, and forming the high performance analysis capability required for large data analysis.

Ling, director of China's Industry Cooperation and Solutions division of Intel (China) Ltd., said that the value after three V was a sufficient complement to the analysis of existing relational data, and that the analysis engine must have the ability to analyze relational and non relational data. Traditional Sans and Nas, he argues, are no longer suitable for large data processing, and are now more in need of scale-out storage architectures, while also requiring real-time data flow processing-that is, high performance analysis capabilities, which require the support of high-performance analysis processors.

Sishing, general manager of Oracle's Greater China region, argues that large data may just be part of an enterprise's complete data-processing platform.

He Yinghua, Director of technical and professional services at NetApp Greater China, presented a distinctive abc--(Analytics), High-bandwidth (banduidth), and content. Big analysis: The analysis is to have an insight into the data, real-time analysis of large data; High bandwidth: Data comes to analyze the results, make a report; Big content: Big content is basically don't lose anything.

He believes that a simple strategy can be taken in developing Hadoop applications with real-time analysis and so on. There are some video aspects to deal with in high bandwidth, and there are some content on the big content. In these three aspects have developed products to be and.

Just need a big practice

Manufacturers have put forward their own solutions, industry users have confirmed the rigid demand for large data. Yang Jing, chief scientist of China Mobile Research Institute, pointed out that the establishment of a flexible vehicle network, can solve the current urban development more traffic jams and other problems, National Gold Securities senior analyst Carey, China Artificial Intelligence Society machine Game Special Committee deputy director Liu Chiqing also put forward the actual needs, such as to analyze the prospects of commercial companies, To solve the go system calculation, and so on.

(Responsible editor: The good of the Legacy)