Data Architecture: Standard Definitions, Challenges, and Optimization Paths

Standard Definition and Scope of Data Architecture

Data architecture is the core of enterprise data management, determining how data is organized, stored, and circulated, providing foundational support for data governance and business decision-making.

Standard Definition

According to ISO/IEC/IEEE 42010:2011 , architecture is the fundamental organization of a system, including components, relationships, and evolution principles. In data management, the national standard DCMM (Data Management Capability Maturity Assessment Model) defines data architecture as: clarifying data requirements through an enterprise data model, guiding the distribution and integration of data assets, and building a data sharing environment.

Core Components

Data architecture typically includes the following core parts:

Data Model – Defines data structures and their relationships, serving as the foundation. For example, an Enterprise Data Model (EDM) ensures different business systems follow unified data definitions, avoiding data silos.

Data Flow & Distribution – Plans how data is transmitted, stored, and used across systems, such as the synchronization mechanism for customer information from CRM to ERP.

Metadata Management – Provides "descriptive information" about data to ensure understandability and consistency, e.g., data catalogs and data lineage analysis.

Data Integration & Sharing – Breaks down data barriers between systems to improve reusability, such as building an enterprise data platform.

Challenges in Data Architecture

1. Data Architecture Not Established as a Management Foundation

Many enterprises only consider data architecture within analytical systems like data warehouses or data lakes, without incorporating it into overall enterprise architecture design.

Data architecture is often misunderstood as merely involving data storage, overlooking the input requirements from business systems.

2. Data Governance Remains Superficial

Overemphasis on data standards without effective mechanisms for data quality and usability.

Lack of unified planning leads to inconsistent data management across departments, preventing synergy.

3. Fragmented Data Management and Severe Data Silos

Business systems store data independently without unified architectural support, causing redundancy and inconsistency.

For instance, a company's customer data resides simultaneously in CRM, financial, and customer service systems, but each system defines data fields differently, making analysis difficult.

4. Over-Reliance on Technology Implementation While Neglecting Architectural Planning

When introducing big data platforms or cloud computing, enterprises often prioritize technology selection over holistic data architecture design.

Data architecture tends to be reactive, receiving attention only when system performance or governance issues arise.

Optimization Paths for Data Architecture

To address these challenges, enterprises need to build a more systematic and stable data architecture to support long-term data governance and business growth.

1. Manage Data from an Enterprise Architecture Perspective

Data architecture should be part of the overall enterprise architecture, not limited to data analysis and storage systems.

During business system design, data inputs, outputs, and circulation paths must be clearly defined to avoid data fragmentation.

2. Build a Stable Data Architecture

Adopt layered design: For example, a "data lake — data warehouse — data service" architecture decouples storage from applications, improving maintainability.

Continuous evolution, avoiding one-time construction: Data architecture should continuously optimize as the business evolves, rather than remaining static after a one-time build.

3. Strengthen Data Governance Collaboration to Ensure Architecture Implementation

The data governance committee should lead the planning and execution of data architecture, ensuring cross-departmental consistency of data standards and architecture.

Data architecture and data governance must advance together, establishing mechanisms such as data quality monitoring and data lifecycle management.