Minsheng Bank opens the way to big data

Recently, China Minsheng Bank Big Data Project officially launched. The project, with IBM and the giant FIR Database Company (SEQUOIADB), sequoiadb to build Low-cost, high-performance, highly reliable and horizontally expanding data platforms for CMBC through IBM biginsights large data solutions and enterprise-class NoSQL databases. To help Minsheng bank through large data analysis to meet the financial industry's big data challenges, to achieve profound industry insights. The platform uses the full integration of IBM Biginsights and partner SEQUOIADB to comprehensively enhance the value of Minsheng banking as a financial leader in the large data age.

Competition drives applications

In fact, Minsheng Bank has put forward "big data" as early as 2010. Minsheng Bank Data Application Center Director Shung recalls, in a technical exchange through his colleagues introduced, he first contact with this concept. By checking the relevant data, Shung feel that large data is far away from them.

But over time everything has changed. 2012, Minsheng Bank of the annual general, Shung found that his leadership actually mentioned important point, to explore business intelligence analysis, as well as large data on how to promote the competitiveness of Minsheng banking business. This to Yuan spring touches very big, after all, although from the Science and technology department angle although with big data have a distance, but from the leadership business level has already put forward the demand.

Under the attention of the senior leaders, the Minsheng Banking Science and technology department began to make the adjustment of the Organization in 2013, and now began to build the data center and set up the data scientist post.

In fact, from the year before last, until this year, from the business sector to the science and technology sector, banking pressure is increasing. Especially from the business front-end to the technical sector challenges.

At present, the whole financial and banking environment is facing many competitive pressures from the industry and even cross-industry. Shung said: "Now the traditional banking in the business of the pressure is very large, including interest rate liberalization, a variety of other sectors of the positive competition and the economic downturn on the risk of banking pressure and so on business pressure." ”

As long as there is a platform, the technical means can be flexible and diverse to practice, so the need for a professional way to quickly put forward, this is a strong demand for the business sector is the biggest challenge. After all, they don't have the knowledge to build a professional (IT).

In fact, in order to enhance the competitiveness of coping with various pressures, Minsheng Bank also takes a number of measures, such as to enhance analytical capacity, to achieve forecasting and insight, including the prediction of a variety of behavior and risk. Then based on the forecast to take measures, how to do marketing, focus on what business, and even what legal means to take, etc., to reduce risk and improve competitiveness.

All this is based on bank data. It includes traditional silver and internet-based data. At present, the traditional data increase is fierce, Minsheng Bank data growth of six months to reach 30%.

Data from the type of e-commerce is equally important to the banking sector. While the bank's business in this area is weakening, this part of the data is becoming increasingly important because of increased involvement of individuals and business customers.

In this part of the user-centered financial behavior and financial status of the more comprehensive data, such as user birth, geographic information, the extension of the trading industry and other data, the bank's analysis of such data more and more attention.

Shung said: "How to make these different types of data to generate business value, large data technology to come." In fact, it is not the impact of the big data on us, but the impact of the business, competition makes the traditional methods are not adapted. ”

Indeed, the financial industry is facing unprecedented technological challenges, the face of a surge of data, how to achieve analysis insights, will be the key to industry innovation and transformation.

As one of the earliest banks in China to deploy large data analysis, Minsheng Bank has affirmed the huge driving effect of data analysis on enterprise value.

IBM's collaboration with the giant FIR database IBM biginsights for enterprise-class strong ease of use and IBM in the financial industry in the field of large data analysis of the rich experience, for the enterprise to achieve customer-focused business development to provide support, but also to build the top of the fine and intelligent application platform laid the foundation.

The fusion of Internet genes

As the first domestic banking group to dabble in large data technology, the Minsheng Bank, in the use of existing NoSQL products (such as MongoDB) in the process of discovering its lack of many enterprise-level functions, such as lack of things and SQL support, and Hadoop system integration is not tight.

Shung also found that large data is an open platform, mostly based on the internet company's open source products, he felt after many practices, banks can not completely copy the original Internet technology model, the new technology platform must be integrated with the traditional industry enterprise-level characteristics. After nearly two years of application system on-line, Minsheng Bank Science and Technology department has accumulated experience, and also set up a specialized team of large data engineers, and actively cultivate a large number of large-scale data reserve forces.

"Technically, large data technology has been in the team of engineers, but the perfect application of large data technology is still a short distance." "The Giant FIR database CTO," said Wang Tao.

With enterprise-class NoSQL database products giant FIR Database, is committed to the Internet has been widely used in the field of NoSQL pushed into the enterprise-level market. SEQUOIADB, through its technical advantages, make up for the shortcomings of most of the existing NoSQL to the enterprise-level function support, and fully meet the demand of Minsheng Bank bank for the next generation large data platform. At present, SEQUOIADB has been widely used in CMBC bank's large data platform.

SEQUOIADB advantage of real-time query, federated query capabilities, for Hadoop strong technical support to achieve a complete integration. Biginsights and sequoiadb good scalability and cost advantage, strengthen the enterprise-level function support, meet the demand of batch analysis and real-time query, make Hadoop function. Make the new integrated data platform solution stand out among the many competing offerings. As a partner, the giant FIR database will work with IBM in more areas to build larger data solutions for different industries.

Wang Tao said: "We can and IBM Biginsights formed very good complementary, such as biginsights more focused analysis, its analytical ability is very strong, our side is the NoSQL database, at the same time can do very real-time data query, in the massive data fast retrieval and data positioning , biginsights from the two sides of an integration, when the formation of perfect integration can be provided to the Minsheng Banking enterprise a 360 degree of any type of business, including real-time access and bulk data analysis. ”

Extending from the case of Minsheng Bank, Wang Tao believes that the current industry challenges focus on both technology and culture. How does traditional business connect with large data? A DBA or business analyst must transform a large data analysis approach with traditional BI analysis.

DBAs actually face two challenges from traditional technical thinking and business thinking. The new big data racks are based on Internet software, such as Hadoop, which is being studied by Yahoo's people; This is the inevitable limitation of traditional thinking for DBAs who value roots in Oracle databases and DB2 database thinking. They have to jump out of the original frame of mind for more imaginative divergence.

The change of mode of thinking is actually the problem of traditional data analyst, and both technological innovation and innovation of thinking mode can meet the needs of the current large data landing.

Application Trend Business Insight

In the past decade, the scale of domestic banks ' assets has reached and surpassed the international level with the rapid economic development. The future development will still face many complicated challenges: from the fundamental change of China's economic growth model, the rise of new customer groups, fierce industry competition and the performance pressure from the bank's own value creation. The financial industry needs to use the analysis and optimization of business to promote its own transformation and innovation.

The rapid growth in Minsheng banking has brought new demands for business insights. Traditional extensive customer marketing strategies are not enough to help banks achieve faster business growth. Minsheng Bank needs to fully integrate customer data, through accurate marketing design to reduce customer churn rate, improve loyalty, with large data technology to different sources of suppliers, customers of the transaction behavior of a comprehensive analysis, to achieve the chain reaction, to build an effective data model to provide customers with a full range of Butler-style non-financial services.

In this cooperation, IBM biginsights based on open source Apache Hadoop's powerful security and ease of use, to help Minsheng bank to improve the transaction flow query analysis system, industrial chain financial management system, as well as private banking product shelf management system.

In response to the huge amount of information in the financial industry 4V Challenge, Biginsights embodies many advantages. For example, with the refinement of business and the increase of enterprise scale, Minsheng Bank's calculation is increasing in order of magnitude. IBM Biginsights can deal with large-scale static raw data analysis, provide multi-node distributed computing, improve data processing capabilities. At the same time, biginsights large-scale parallel linear scalability ability, can deal with massive text processing.

Minsheng has been able to improve its data analysis capabilities since its deployment of the Hadoop computing platform in 2012, but is still constrained by the development cycle in processing diversity data. IBM Biginsights integrates a robust, scalable, structured and semi-structured descriptive language jaql that handles various types of data analysis and enables comprehensive data analysis in different application scenarios. In particular, the built-in semantic analysis of the text in the cluster, for a variety of sources of text to provide high-performance processing, tagging and analysis functions.