Quenching the body and returning to the element, operator resource domain metadata management

Resource metadata is the basic supporting element for the development of resource management business in the communication industry. It defines related concepts, relationships and rules in the field of resource business, that is, the specifications, attributes, and dictionaries of network resources such as various facilities, cables, equipment, and links. And related storage model and other information. High-quality metadata is a necessary condition for improving business efficiency and strengthening management and analysis capabilities. Building a resource metadata data system is a key basic work for improving the quality of resource data.

During years of business development, operators have gradually accumulated the following problems in the management and use of resource domain metadata:

Different standards, loose system

• Insufficient abstraction and induction, different definitions of division of resource specifications, and uneven granularity of specifications

• The business definitions of specifications and attributes are imprecise, and similar meanings exist

• Compared with industry norms, there is a certain degree of difference

Insufficient aggregation mechanism is not conducive to improving and improving metadata

• Metadata gradually loses connection with the implementation of the project, forming an information island

• There is independent evolution of metadata, lack of convergence and convergence, and it is difficult to guarantee the comprehensiveness of product metadata

Weak analytical skills, difficult to find problems

• Lack of metadata analysis capabilities, lack of visual exploration means for data relationships in metadata, and data opacity

• Lack of quality audit methods, abnormal metadata configuration information can easily cause business function problems

Value mining is weak and data realization is difficult

• In addition to defining the basic usage of business data, the use value of metadata in business functions, data services, data quality, data migration, etc. has yet to be explored

A set of metadata with a stable foundation, a complete system, and accurate data is the prerequisite for the stable operation of the business system, and the governance of metadata is a normal task.

This article will introduce the management ideas of resource domain metadata from the following four aspects.

Build a metadata system with a pulse to follow

The construction of a standard resource metadata system will fully turn the operator's network resource data into digital assets. Through the use, processing and analysis of resource data, it will provide support for business operation and development, and ensure that the operator's data assets can play their full potential. business value.

Based on the industry norms, combined with the existing metadata, standardization is carried out with the goal of unifying business definitions, unifying physical models, and unifying information items to form a resource metadata standard system.

Metadata System Construction Process

Sorting out the specifications in the metadata system : make top-level abstractions from points, lines, networks, and signs, and decompose and define the major categories, subcategories, and specification systems of business objects step by step, and incorporate carrier network resources into specific locations in the system to cover all resources Application requirements.

Sorting out business attributes in the metadata system : Sorting out standard terms for various business attributes, such as equipment name, equipment code, placement location, equipment capacity, equipment model, etc., cataloging and standardizing these business terms. And map the business attributes to the corresponding physical library table fields, including field names, field codes, field value ranges, whether they are primary keys, whether they are unique, whether they are not empty, etc.

Modeling in the metadata system : use the idea of ​​category modeling to maintain the stability of the main table of the category, and realize flexible expansion by inheriting the category and adding the extension table. The main table + extended table + vertical table is used to store business objects through metadata encapsulation, and the relationship between specifications is concentrated on the main table of large categories to ensure the balance of model convergence and scalability.

Metadata system (category-subcategory-specification-attribute)

The product metadata system, from the aspects of specifications and attributes, takes into account common and individual needs, and realizes coexistence management. At the specification level, there are standard (standard) specifications, product (extended) specifications, and (project) private specifications; at the attribute level, standard attributes, product attributes, and project private attributes are distinguished. Establish and form a robust metadata system (category-subcategory-specification-attribute) with industry standards as the core and product scalability.

Category : with points, lines, nets, and logos as the top layer, geometric shapes and functions are used as the basis for category division to form a category entity of network resources.

Subcategories : According to the functions and characteristics, further stabilize the classification of large categories, and use them to manage business objects in a directory to improve management efficiency and reduce management difficulty.

Metadata category-subcategory division

Specifications : Specific business objects, content that expresses the specific characteristics of the business, and systematically manage according to the sub-categories. In the specification definition process, in addition to the existing standard specifications of the industry norms, the product expands and supplements the specification definitions of relevant business objects within the framework of the large-category-subcategory system according to the needs, so as to meet the needs of production management.

Attribute : The business attribute of the business object. It is divided into three categories: standard attributes, product attributes, and private attributes. Standard attributes are attributes that comply with industry norms, product attributes are attributes that do not belong to industry norms but have certain generality, and private attributes belong to a project. Business attributes will be stored in different tables (main table, extended table, vertical table) according to attribute classification and nature.

Gather and improve metadata, gather sand into a tower

Resource domain metadata, which is the core data of the industry, is scattered in various production systems. It is necessary to collect, clean and integrate the scattered metadata to complete the convergence and continuously improve the core assets of product metadata.

Aggregation and improvement of metadata process

Acquisition : Obtain existing metadata from different data sources, including data such as specifications, attributes, and dictionary values;

Cleaning : analyze the collected metadata, judge the specifications, attributes, dictionary values, etc. according to some business principles (whether it is in use, whether it is reasonable, whether to log out, etc.), clean the collected metadata, and remove the false and save the essence of the metadata. Eliminate invalid/useless data;

Fusion : According to the principle of metadata division and definition, analyze and fuse the cleaned specifications, attributes and other data, including identifying similar specifications for merging, identifying and redefining unreasonable specifications, identifying attributes with similar meanings for unified definition, and Meaning dictionary types and dictionary values ​​are merged and unified;

Aggregation : Through the continuous aggregation process, the completeness and comprehensiveness of metadata are realized, and finally a unified and comprehensive metadata system for products is formed.

Metadata management analysis, insightful

As mentioned at the beginning, high-quality metadata is a necessary condition for improving business efficiency and strengthening management and analysis capabilities. Through unified management of metadata, visualization of metadata assets, and analysis of data relationships, product metadata can be continuously improved and guaranteed.

Metadata Management Analysis

Asset view, a complete picture of data at a glance

The metadata asset view organizes information from the macro level, and strives to synthesize information assets from a management perspective, display the macro information of assets globally, and effectively tap the potential value of information.

The asset view shows an overview of the metadata of products and items:

Overall data overview : display the number of specifications, attributes, and references of key specifications for products and projects; display the classification and proportion of specifications, including the proportion of standard specifications, proportion of product specifications, and proportion of private specifications;

Configuration horizontal analysis results : display the results of horizontal comparison between the specifications and attributes of the project landing and product metadata, and show the configuration consistency of specification attributes;

Data quality evaluation : Based on the metadata audit normal number, standard specification usage rate, standard attribute usage rate and other information, qualitatively evaluate the metadata situation, discover shortcomings in metadata configuration and optimize content, and promote the overall quality of metadata.

Lineage analysis, exploring data relationships

Lineage analysis allows users to understand the relationship between metadata, the specific content of each relationship, and what kind of output to produce according to their needs. Starting from a certain resource specification, several internal relationships of its metadata are traced in depth:

• Analyze the reference relationship between specifications and projects, grasp the usage of specifications by various projects, and compare and analyze the configuration differences of specification attributes in projects with products;

• Analyze the relationship between specifications, discover the absence of relationships between specifications, and identify high-frequency specifications cited by relationships in metadata;

• Analyze the reference relationship between specification attributes, business tags, and APIs, mine the association between attributes and APIs, and judge the impact of attributes on business logic, providing important references for attribute configuration adjustments.

Quality management to improve data accuracy

The nature of metadata determines that its quality has a wide range of influence. There are many business and technical configuration information items in metadata, and many information items are key configuration information. These information either affect the quality of instance data or affect the system. Functional logic, through data auditing methods, can detect problems in time and ensure the accuracy of related metadata configuration.

Metadata auditing is the process of checking the legality of metadata itself, and checks data attributes, data attribute relationships, and data table relationships through audit rules and execution audits.

Metadata configuration problems are found through metadata audit rules, and clear exception repair prompts are provided to facilitate managers to quickly repair metadata problems through management capabilities.

Examples of metadata audit rules

The value of metadata is reflected, and the "number" is used to the fullest

A stable and unified metadata system is conducive to the unification of the platform service layer and the core data layer. On this basis, a unified version of the production system is formed to support the application construction of the upper-level planning, construction, operation, maintenance and optimization. Here we will discuss the use value of some high-quality metadata, and believe that more possibilities can be found in the actual application process.

Basic function metadata driver

The basic maintenance functions of resources (query, addition, modification, etc.) can be quickly implemented based on metadata drivers. The specification attribute is the display of controls that require an interface. The UI forms of the attributes are different. There are various input elements such as text boxes, drop-down boxes, date boxes, and large text boxes. By referring to the HandleBars semantic template library, you can Generate a front-end view to quickly generate a web template for a business object management interface, encapsulate this whole set of implementation mechanisms to form basic capabilities such as attribute components and query components of the product, and use metadata configuration to simplify the development of business functions through attributes, query and other basic components .

Metadata-driven generation of basic functions

The conversion from requirements to functions requires only two steps: first, analyze requirements and organize metadata, and configure through metadata management capabilities; then, business-side applications load configuration data through components, and automatically generate functional interfaces for query, addition, and modification. Get rid of the traditional mode of hard code implementation, so that the basic management functions of business objects can be realized flexibly and quickly.

Data service metadata driver

Based on the capability sharing requirements of resource data, it is necessary to provide unified standard external data services and business service capabilities. Driven by metadata, CRUD basic data services oriented to resource specifications can be quickly provided through configuration means.

The definition of data service, the attribute selection of each entity specification required in the data service, and the attributes required to determine the input parameters, output parameters, and association conditions of the data service.

Selection of Specification Attributes Involved in Data Services

The logic assembly in the data service sets the association conditions in the data service:

• Establish the association relationship of attributes between specifications, and define constraints such as external associations for association conditions. As shown in the following figures ① and ②, establish the association relationship between attribute A of specification A and attribute A of specification C with "=" to form condition 1; assemble the two attributes of ③ and ④ to form condition 2; according to logical requirements, Can continue to form the associated condition N;

• Then, condition 1, condition 2, and condition N are further assembled according to the logical needs of the data service (combination of and, or, etc.);

• The assembled conditions will eventually be translated into the association of table fields at the database level according to the metadata attribute storage configuration.

Conditional assembly of data services

Generate data services driven by metadata, configure the relationship between related resource specifications and business attributes according to business requirements, reduce the difficulty of data service development, and easily guarantee the standardization of data services.