Recently, Gartner clearly pointed out in the research report that "metadata management will be the core infrastructure of enterprise informatization in the future". Indeed, in the big data environment, if enterprises do not manage all kinds of complex information through metadata management, it will be difficult to make effective use of information. However, many enterprises gradually find that the value that metadata management directly brings to enterprise business innovation is very limited. What is the current situation of metadata management? How to fully release the commercial value of metadata management? What practical experience can be used for reference?

Directory:

First, the status analysis: lonely enterprise metadata management

Second, the solution: release the metadata value for the business.

Third, technical practice: Puyuan's enterprise metadata management practice.

First, the status analysis: lonely enterprise metadata management

The main reason why metadata management cannot bring direct value to business innovation is that the current metadata management is too "lonely", which is mainly reflected in the following four aspects:

Narrow scope of management: At present, the narrow scope of metadata management is mainly reflected in two aspects: first, only metadata related to data warehouse is managed, and metadata management does not extend to the whole enterprise level; Second, even in the field of data warehouse, only part of technical metadata is managed. The lack of technical metadata leads to the failure of metadata system to show the real situation of enterprise system, and the lack of business metadata leads to a pile of table structures lacking specific business meanings.

Business is difficult to integrate: due to the lack of integration with business, many enterprises, after completing the metadata management system, find that only people in the data management department are using it, and people in other departments simply don't use it. All systems should serve the business, and the system used by only one department is lifeless. How to make people in all departments use the metadata management system is a very important thing.

Lack of application scenarios: At present, the application scenarios of metadata are mostly limited to the internal functions of the metadata system, such as consanguinity analysis, pedigree analysis, version management, etc. These functions are used by enterprises in informatization, but in fact, the role of metadata is not only reflected in these limited functions.

Imperfect technology: Imperfect technology leads to the above series of problems. The scalability is not strong, resulting in many metadata management tools can not manage enterprise business data and architecture; Poor acquisition ability leads to manual auxiliary recording, high labor cost and inability to establish a complete information link; Real-time performance is not high, which leads to the fact that many enterprises' metadata management still stays at the stage of T+ 1 (not even T+ 1), so they can't know the status of data assets in real time and can't keep up with the growth rate of enterprise data.

Second, the solution: release the metadata value for the business.

In order to get rid of the above-mentioned "loneliness" situation, metadata management can start from four aspects:

Enhance scalability in technology, realize automatic acquisition first, and enhance scalability. At present, most metadata products conform to CWM standards, and can only manage metadata related to data warehouse, but do not support the collection of metadata such as business and architecture. Based on the MOF located at the bottom of CWM, the scalability problem can be effectively solved. In this way, Puyuan uses metadata to manage the whole micro-service architecture in The Platform, a new generation platform, and realizes the unified management of services, data and operation and maintenance processes. Second, improve the collection ability. Metadata management needs to collect all kinds of metadata, including scripts, stored procedures, reports, etc. Moreover, it is necessary to analyze various syntax trees, which requires a high collection ability of the system. Puyuan's metadata management platform has realized automation to improve the collection performance of the system, and can collect a variety of metadata, and the collection rate is almost 100%.

The premise of improving business metadata management ability, making it widely used by business users and fully releasing the business value of metadata management is to let business personnel learn to use metadata management system, so how to collect business metadata becomes very important. Collecting business metadata through Excel requires business personnel to fill in various data items. Even if it is collected in the system, the association between business metadata and technical metadata cannot be established, and the technical path of business data cannot be reflected. In order to solve the above problems, automatic acquisition is essential. In addition, by analyzing business metadata from documents and then corresponding technical metadata, more users can participate in the use of metadata system.

At present, the metadata management system is limited to using specific internal functions. In order to give full play to the value of metadata management, metadata should be made into a service-oriented system, embedded in various scenarios of enterprises, providing capabilities for other systems, making other systems more automated, making metadata management a very important infrastructure of enterprises and integrating into all aspects of enterprise informatization.

Open the data production line to form self-service data supply. Many enterprises have encountered such a problem: when developers ask the data department for data, the data department needs to explain the meaning of the data one by one and guide the data to the past. The difficulty of interpretation increases with the increase of the system. Metadata can help enterprises to build self-service data production lines, so that developers can easily obtain the required data from the data platform, make the whole industry line simpler, reduce the workload of data departments and improve the satisfaction of business developers.

Third, technical exercises:

Practice of metadata management in Puyuan enterprise

China Eastern Airlines: The original data map of business data can only be understood by technicians, but business personnel are more concerned with business-related matters and the distribution of business data. Therefore, Puyuan combed the whole model of China Eastern Airlines and presented these data in a way that business personnel can understand. Through the construction of data asset management platform, China Eastern Airlines realized the presentation of operational data map. Drawing on the experience of Delta Air Lines, Puyuan analyzed nearly 2,000 entities in the aviation domain model, checked the data of 1249 business system tables one by one, and combed out a three-tier data map including data subject domain, data entity and business system: 13 subject domains including flights and ticketing, and provided up to 227 business entity directories and definitions for each subject domain.

Zhejiang Electric Power: Self-service data query for business users based on business metadata. Due to the limited IT level of business personnel, they can't design their own reports or self-service data query, so Zhejiang Electric Power has to deal with a large number of demands from business personnel every day. Puyuan helped Zhejiang Electric Power to sort out all the metadata and integrate it with the reporting bank. Through the corresponding business metadata and technical metadata, business personnel can query the technical channels corresponding to business data, thus designing their own reports, greatly reducing the workload of data departments; Puyuan also helped Zhejiang Electric Power to make a data map of business process. Business personnel can directly see the corresponding data of each process node from the data graph. In this way, business users can find the required data by themselves.

Debon Logistics: Automatic real-time data asset collection Debon has established a relatively perfect data platform-universal metadata management platform, which has helped Debon manage more than 90 business systems, optimized the entire existing process, and realized automatic collection of various environments (databases, ETL, services, reports, GP, stored procedures, etc.). ), the accuracy rate is 95%-99%. After automatic collection, Debon safely eliminated 50 useless reports unrelated to any system from more than 500 reports, greatly reducing the maintenance cost of reports.

For the test deployment stage in development, under ideal conditions, the metadata in design state, test state and production state are consistent. Usually, the development manager needs to compare the design state with the production state, and judge whether the system can go online according to the comparison results. Operation and maintenance personnel also need to analyze the impact on other systems before the system goes online. Before that, all these things need to be done manually. After the completion of the project, metadata has been integrated into every development link of Debon. Through the metadata management system, the metadata differences between different states can be completely compared, and the comparison difference report can be directly given. The development manager and operation and maintenance personnel can judge whether the system can go online according to the report.