Abstract：When an enterprice implements a CMDB configuration management tool and impots all relevant instance data into the platform, it only completes the first step in CMDB construction. The second step is to consider how to leverage this data and generate value from it. During this process, managers need to address these critical questions: At this stage, data governance becomes paramount.
Keywords：CMD, Data governance, Data security
01 Why Data Goverance Matters
Credibility & Accuracy: It enables institutions to ensure the accuracy, intergrity, consistency, and timeliness of data by establishing data quality standards and audit rules, as well as monitoring and addressing data quality issues.
Compliance & Security: Data goverance ensures that insitutions adhere to the industry standards and regualtion in data goverance and usage. Through the implementation of data privacy and security policies, along with controlled data access and permissions, insitutions can reduce the risk of data breaches and non-complant operations.
Consistency & Integration: It guarantees consistency and integration of data across different systems and departments within an institution. By defining data interaction standards and specifications, and establishing mechanisms for data integration and sharing, data governance eliminates redundancies and inconsistencies, thereby enhancing data value and utilization efficiency.
Discoverability & Accessibility: Data goverance helps insititutions better manage and organize data resources, making them easier to discover, access and ultilize. By establishing a data catalog and metadata management system, along with providing corresponding data query and analysis tools, it fosters data-driven decision-making and innovation.
Ownership & Accountability: Data goverance clarifies data ownership and accountability, ensuring lawful and appropriate usage. By establishing data governance processes, and by defining governance rules, it can clearly define data ownership and usage permissions, reducing the risk of data abuse and misuse.
02 How Leading Enterprices Approach Data Goverance
Let’s take a client in the insurance industry as an example. This client has been developing its CMDB for about 3 years. The underlying data collection has largely entered the mid-to-late stage, and the data volume is already considerable. As the project progressed, the client proactively initiated a series of data governance efforts around CMDB data, which have delivered meaningful and tangible results.
2.1 Analysis of the Current CMDB Implementation
(1) Comprehensive Implementation
The CMDB module, as part of the overall project, has been develpoed for nearly three years and has completed two phases. The first phase primarily focused on foundational enablement, putting the core configuration management capabilities into use. The second phase shifted its focus toward the business side, consolidating and streamlining CMDB models and field information to align the actual business scenarios. To date, there are 108 valid models, approximately 280,000 instance records, 737 business entities, and the management of data center infrastructure covering wind, fire, water, and power systems.
(2) Strong Leadership Support
Currenttly, the leadership strongly recognize that CMDB serves as the foundation of the platform, with all scenario development built upon it. With strong support from leadership, the CMDB implementation team has been able to effectively mobilize other business units and subsidiaries, driving broad participation and strong collaboration in CMDB construction and improvement.
(3) Focus on Data Goverance
With the volume of foundational data increased, the project team has begun to place greater emphasis on data governance while continuing to expand upper-layer business scenarios. Adhering to the principle of "data ownership and accountability," the team defined a domain-based configuration data matrix for CMDB assets and clearly assigned primary ownership for model management.
2.2 CMDB Data Governance Approach
(1) Current Issues
Challenge: Once CMDB construction begins, institutions often encounter a large volume of communication overhead and various issues. As trivial tasks accumulate, it becomes impossible to manage every detail perfectly. Consequently, while problems are widely recognized, they often remain unresolved, hindering tangible implementation.
Solution: Based on the current state of CMDB, prominent issues are categorized, such as those requiring leadership's intervention, data quality problems requiring audits, and other critical dimensions. These issues are reported through weekly meetings. This informs the senior leaders about current challenges and support needs, while weekly updates ensure the issue log remains current and actionable.
Sample Issues List:
(2) Governing Incremental Data
Challenge: As the importance of CMDB increases, an increasing volume of data flows into the platform. Effective management of incremental data is critical. Without proper controls, data quality quickly deteriorates, making it difficult or risky for downstream to use the data. Common issues include unannounced addition or removal of field information, inconsistent data entry requirements across different periods, and the need for repeated manual validation.
Solution: Integration with external systems, development of collection plugins, and APIs are employed to increase data collection rates. For example, linking to an architecture management platform for business information or to Huawei Cloud for cloud resource data. A monthly integration plan with clear milestones was established and tracked. On the other hand, CMDB data governance is tightly integrated with workflow and ticketing processes. Physical asset provisioning and decommissioning is governed through process workflows and tickets, with multi-level approvals, ensuring that every data record entering the CMDB is traceable and verifiable.
Sample Integration Plan:
(3) Activating Existing Data
Challenge: The CMDB contains invalid fields and models that have not been used or updated for a long time. This has resulted in bloated models with limited business value, a large number of models without interconnected consumption scenarios, and a heavy maintenance workload. Operations teams often is stuck in a cycle of daily manual data corrections, with no reference standards or clear criteria to determine when the data can truly be considered corrected.
Solution: To ensure the accuracy of existing instance data, two key measures are implemented. First, the audit capabilities of the configuration management center are fully leveraged to conduct periodic audits. Non-compliant data was exported and assigned to the responsible model owners for correction. Audit results are then communicated to internal stakeholders via operational weekly reports. Secondly, business scenarios are developed to actively using underlying CMDB data. Simultaneously, integrate with surrounding external systems to synchronize data through event subscriptions and APIs, expanding scenario coverage and ensuring data remains dynamic and flowing.
Sample Existing Data Governance Case:
(4) Indentifying Business Scenarios
Challenge: Once CMDB data is in place, a key challenge becomes how to effectively utilize it and demonstrate its value of the data, especially investing annual labor costs. These are the most challenging issues for every CMDB manager.
Solution: Business consumption scenarios are systematically identified, along with the models and instance attributes consumed by each scenario. Comparative analysis is then used to optimize models and attributes by eliminating obsolete or rarely used attributes to ensure every field’s relevance. In addition, regularly publish access statistics for each business scenario to evaluate their current value and impact.
Sample CMDB Consumption Scenario:
(5) Establishing Resposibility Matrix
Challenge: As CMDB development advances, the number of models increases accordingly, and the needs for adjusting and maintaining each model begin to diverge. With limited capacity, a single administrator cannot always meet the demands of business departments, leading to conflicts, misunderstandings, and loss of control. This process also spent a lot of time for cross-departmental communication.
Solution: A responsibility matrix is established to clearly identify the owner of each model, adhering to the principle of "data ownership and accountability." By assigning data governance responsibilities to specific individuals, each model can be managed consistently throughout its lifecycle.
Sample CMDB Responsibility Matrix:
03 Achievements of Data Goverance
3.1 Establishing the Core Role of CMDB Source of Data
The managed assets are diverse, extensive and comprehensive, ranging from upper-layer business systems down to underlying devices, data centers, and racks. The CMDB continously integrates with various asset systems, acquiring corresponding data back into the CMDB. At present, CMDB already provides source data for applications such as the alarm center, O&M processes, visualization dashboards, scenario-based applications, mobile O&M, and asset management. As its integration with external applications advamces, its core position has become irreplaceable.
3.2 Tangible Results of Scenario-Based Construction
With abundant raw data, a wide variety of data-driven consumption scenarios have emerged, such as displaying queries from both management and operational sides, self-service queries that facilitate cybersecurity, rich instance information to improve alert-handling efficiency, automatic feedback on equipment installation and decommissioning processes, real-time synchronization of data for asset management applications, automated deployment of firewall policies, execution of batch script operations based on the CMDB, and regular batch generation of offline QR code.
3.3 Rapid Development of O&M Automation
As the scope of managed services and the associated node information continues to expand, the number of nodes handled by an individual has become massive. Relying on manual processing is no longer suitable for current needs. Therefore, automation of operations has become inevitable. Leveraging the foundational data from the CMDB to achieve automated operations not only reduces labor costs and improves daily O&M efficiency, but also enables rapid problem detection and resolution, thereby ensuring the secure operation of data centers.
3.4 Shifting Daily O&M Mindsets Toward Rational Management
Daily operations often focus on completing tasks in a routine, with little reflection on existing issues or how optimizations could lead to better experiences and higher efficiency. When urgent incidents occur, responses may be chaotic. Therefore, changing this mindset is crucial. By rationally analyzing current problems, exploring solutions, and devising strategies to prevent recurrence is the development of management. Furthermore, with the development of scenario-based applications, this approach can further explore new value in operation.
04 Data Goverance Tool Support
Data goverance can only achieve its great effect through the combination of huaman expertise and tools. CanWay CMDB has integrated data governance capabilities, adhering to the philosophy of "top-down design, clear accountability and authorities, data-driven decision-making, and closed-loop assurance". It provides Out-of-the-box data governance capabilities and can monitor trends in data quality of CMDB.
[Operational Analysis - Operational Audit Rules]
Function: To activate existing data, operational audit rules can be set to review the stored data. These rules periodically screen for non-compliant data and generate corresponding audit results. Through audit rules , invalid or neglected field information can be identified and analyzed.
[Operational Analysis -Quality Operation Dashboard]
Function: The Quality Operations Dashboard provides real-time results generated from the execution of Operational Audit Rules. It offers a clear view of audit results across various modules, facilitating further data analysis and mining. This helps identify the data maintenance behaviors of underlying users and provides a data-driven basis for work planning and corresponding strategies.
[Operational Analysis - Quality Operation Dashboard - Execution Results]
Function: For the audited results, the system will carry out the data supplementation and correction according to accountability matrix based on the principle of "whoever is responsible claims ownership." This assists operations personnel in understanding the objectives and tasks, promoteing collaboration, and facilitating understanding and analysis.
