Interview EMC Liuweiguang: Greenplum's Big Data strategy

Big data has huge room for development, as evidenced by governments ' policies on big data. U.S. President Barack Obama's big data development plan brings together the nation's top experts to turn data into business assets and values. China's big data market is also no doubt that big data contains huge potential and business opportunities. In recent years, it vendors have launched their own large data strategy, EMC is no exception.

EMC, as a traditional storage vendor, has been a leader in global storage. Recently, IT168 reporter to EMC Data Computing Products division general manager of the Greater China, Mr. Liuweiguang interviewed, together to explore the large data business value and large data market space, and a series of topics, and in-depth analysis of the Emcgreenplum large data strategy.

According to Liuweiguang, he joined the EMC company in November 2011, and was responsible for the overall operation of the Greater China area of the Data Computing division. Previously worked in Oracle for nearly eight years, starting with the system architect, and then the senior manager of the Greater China Enterprise Architecture team, and became the director of the Exadata China Product division in 2010. Liuweiguang's EMC Data Computing Product Division was formerly an EMC acquired Greenplum company. Greenplum is a global leader in distributed databases, and EMC buys Greenplum companies by valuing its future business value, especially in the great potential of the big data market. In the global establishment of independent operation of the Division, namely, Data Computing division. Liuweiguang, as the manager of the Greater China area of the division, is mainly responsible for the overall business promotion of the product in China, including sales, pre-sales, service, brand promotion and so on.

Three milestones in database development

When it comes to the development of database applications, Liuweiguang said that the 15-year history of it construction in China's large enterprises has roughly three milestones: the first milestone is a relational database oriented to the transaction type, and in the late 90 to the beginning of 2000, China's IT system development is the fastest, the most extensive construction, the largest investment in the years. China's it construction from a relatively backward stage, towards a new development process, more and more foreign enterprises into China. China's IT system construction, especially in the telecommunications, banking and government sector has undergone a large-scale transformation. Traditional databases are typically OLTP, that is, transaction-oriented database, usually supporting the basic business functions of the system operation and the most basic information needs of enterprises, telecommunications industry construction billing system, CRM system, customer service system, banking industry to build the front core system, other industries to build basic customer management, Marketing System. The main function of this system is to store data, to provide some services to customers, just as to solve the problem of food and clothing for people's lives, this database is to support online transaction processing database.

The second milestone is the Data warehouse, which can be likened to a higher level of demand for it construction to address food and clothing problems. At this stage, the enterprise's IT operations to reach a certain level, accumulated a lot of experience. The enterprise discovers the data is the very important asset, but does not have the perennial data to become the Technical Foundation which directs the enterprise operation. At this stage, many large enterprises began to build data warehouses. The predecessor of Data Warehouse is analysis report system, that is, the data is extracted from the database to form the statistic report, but this report usually does not instruct enterprise operation and decision analysis. To the Data warehouse stage, the data not only form the report, but also according to various themes, the internal needs of the enterprise processing, analysis, and then form a decision support data source. After storing the data, mining the data, processing the data, displaying the data, the result of the data becomes the most important technology input for the enterprise's next operation and the development of the market strategy.

The third milestone is large data. Cloud computing has been raging in recent years, and cloud computing and big data are mutually reinforcing relationships in many ways. At this stage, with the impact of new technologies and technological means of innovation, as well as the impact of Internet technology on the development of IT industry has become increasingly apparent. The emergence of cloud computing has a huge challenge to the data warehouse, and how to deal with the data that the traditional relational database cannot handle is the biggest challenge for the new technology. Mass data and large data are different concepts, and mass data usually refers to the collection of data stored in the traditional relational database after the database table structure is designed and processed. Large data is also larger in data capacity than mass data. In addition, large data sources are rich in data types, including unstructured and semi-structured data from the high level of information from the Internet and traditional enterprises, as well as the ever-evolving historical archive data, which is far from being quickly loaded by current technologies, And it is not the traditional database and data Warehouse storage management and analysis.

Large data contains commercial value

Now a lot of companies are talking about big data, experts on large data have their own views, liuweiguang that large data mainly includes four features: first, large data data volume is very large; second, large data have very complex data sources; third, large data has very complex data structure, It is not a collection of data that the traditional relational database can handle; four, the effect of large data is very low, that is, the value of processing data in the unit time is relatively low, but if the rapid processing and analysis of large data in unit time, it will produce unexpected commercial value.

Although the effectiveness of large data is very low, often require a large number of computing power, but the large amount of data can not be underestimated commercial value, Liuweiguang very optimistic about large data market development prospects. First of all, from the point of view of commercial value, in the financial sector, through the mining and analysis of various customer transaction information, can be through decision-making analysis and maximize the increase in corporate sales profits. From a positive point of view, this way can find High-value customers, the corresponding products to carry out accurate marketing, from the negative point of view, can also be carried out fraud analysis, reduce the operational risk of enterprises.

In the field of telecommunications, the most common application of large data is through the analysis of user signaling data for roaming users to send welcome text messages. With the development of large data analysis in telecom industry, the analysis of user behavior data for accurate marketing will gradually become the new business type of operators. This precision marketing embodies a new feature of large data: low value conversion rate, which is to analyze a large amount of data in a very short time, to provide valuable value-added services for everyone, mining potential business opportunities. In addition to the traditional telecommunications and financial industry, scientific research institutions in the field of large data applications also have broad prospects, scientific research institutions to take out the dust-laden data to use new technology for mining analysis, solve the problems of scientific research work.

(Responsible editor: The good of the Legacy)