How Unicom Digital’s Integrated Data Platform Revolutionizes Metadata Management

Background and Challenges

In recent years, China has elevated data to a strategic resource, defining it as a foundational national asset and a key production factor. The establishment of the National Data Administration and the three‑year Data Element Action Plan underscore the growing emphasis on data’s economic value and the need for robust data management across massive market scales, diverse applications, and extensive data assets.

Broadly, data includes any recorded information, which after collection, governance, and integration becomes a "data resource". When such resources generate economic benefits or have measurable costs, they become "data assets" suitable for accounting and trading. The market is currently in the data asset management phase, requiring solid foundations for future asset accounting and transactions.

Unicom Digital’s integrated data platform has earned multiple recognitions, including awards from the Ministry of Industry and Information Technology, the Ministry of Ecology and Environment, and DAMA. In the 2023 Data Governance Industry Map 2.0, twelve of its capabilities were featured, and its metadata management module was among the first to pass a specialized evaluation by the China Academy of Information and Communications Technology.

Metadata Management Challenges and Solutions

Key challenges include:

Operational difficulty: Users struggle to locate, understand, and operate metadata, often requiring manual effort.

Management difficulty: Integrating multi‑source resources, controlling storage consumption, and deriving value from metadata are complex.

Unicom Digital addresses these with three approaches:

Automation : Minimal human intervention is needed after deployment; the platform runs autonomously and only requires manual checks when issues arise.

Linking+ : The platform connects internal modules and external management systems via standardized interfaces, enabling unified data management.

Intelligent : Leveraging large models and data virtualization, the platform becomes smarter and more user‑friendly.

Exploration and Practice

The integrated platform manages over 500 internal databases, 23,000 tables, 2 million fields, and 2,000 data nodes, handling daily data volumes of around 500 TB. It serves more than 1,000 enterprise customers and has delivered over 50 external projects, including more than 20 provincial‑level initiatives.

The metadata management workflow covers three metadata types—business, technical, and management:

Business metadata : Establish data standards and build logical models.

Technical metadata : Create physical models, capture data collection, processing, and warehouse activities, and store quality checks, API definitions, and other technical artifacts.

Management metadata : Integrate lifecycle management, operation logs, and data security classification.

The platform automates metadata collection, publishing, versioning, and permission control, and performs quality validation through consistency, completeness, and standard‑coverage checks. Statistical analysis modules provide quality reports and resource usage insights.

Full‑chain lineage management, enabled by the "Linking+" capability, tracks data flow across more than ten layers, supporting use‑case scenarios such as application issue localization, impact analysis, usage‑complexity ranking, and data‑island detection.

Data responsibility is enforced by linking metadata to an accountability system that assigns owners, technical leads, and business leads, ensuring clear responsibility for data issues. Lifecycle management classifies data into hot, warm, and cold tiers, automatically moving or deleting data based on usage patterns.

Summary and Outlook

Future directions focus on metadata intelligence:

Intelligent classification : Automatically assign domain and layer tags using business, management, and sample data combined with lineage information.

Intelligent completion : Use AIGC techniques to auto‑fill missing metadata during collection.

Intelligent recommendation : Suggest relevant metadata during search and recommend responsible owners during data accountability processes.

Metadata management also supports data‑asset accounting by providing comprehensive asset inventories, cost measurement (storage, compute, labor), and quality‑based valuation, thereby bridging technical governance with financial assessment.

Q&A

Q1: How can B2B businesses break data silos while complying with regulations that prohibit plain‑text data exchange?

A1: Adopt a trusted data‑resource space model that enables privacy‑preserving computation or data sandboxing for regulated data sharing.

Q2: How should customers with their own platforms manage metadata?

A2: Define standardized interfaces and meta‑models to integrate external platforms with minimal development effort, facilitating unified metadata management.

Q3: What is the scope of full‑chain lineage, and how does post‑warehouse data contribute to it?

A3: Once data enters the integrated platform, it falls under full‑chain lineage monitoring, covering tables, indicators, tags, etc. Post‑warehouse data requires standardized permission management and upstream/downstream integration for comprehensive tracking.