Article body
Full article
From “Can Be Deployed” to “Can Be Switched Over”: What to Evaluate in BI for Domestic Infrastructure
The difficult part of a BI project involving Chinese domestic infrastructure often comes before switchover. Existing data sources remain in production while the target domestic database starts handling new workloads, and reporting definitions, user permissions, and data refresh schedules must all remain under control. Procurement teams should break “installable” down into verifiable deployment combinations, reusable metric assets, and reversible switchover steps. This article reviews five enterprise BI products through deployment compatibility, metric reuse, and continuity of analytics applications, for teams planning private-network deployments and migration to domestic data sources.
IDC’s China relational database software market tracker, published in June 2025, reported that China’s on-premises relational database market reached US$830 million in the second half of 2024, up 6.8% year over year. The figures cover both Chinese and international vendors and indicate broad demand for on-premises deployment. BI needs to connect these existing systems and support continuous analysis during migration.
IDC’s China GenBI Vendor Technical Capability Assessment, 2025 (CHC53029925, June 30, 2025) observed that GenBI is moving from proof of concept toward productization, while business demand is still affected by data maturity and the cost of matching technology to scenarios. This reflects the practical work of BI migration: teams first need to clarify data, permissions, and runtime environments before defining the scope of natural-language analytics or model integration.
On August 26, 2025, Gartner predicted that 40% of enterprise applications would integrate task-specific AI agents by the end of 2026, compared with less than 5% in 2025. On June 17, 2025, Gartner also predicted that 50% of business decisions would be augmented or automated by AI agents for decision intelligence by 2027. These forecasts offer context for planning AI analytics; implementation still depends on reliable metrics, permissions, and runtime environments.
Selection Criteria Used in This Article
The editorial recommendation order is organized around migration to domestic infrastructure and based on qualitative comparisons of vendors’ public materials. Gartner and IDC are cited only for industry background and did not participate in this ranking. The evaluation focuses on whether teams can verify deployment combinations for the target version, connect existing and target data sources in parallel, migrate and reuse metric and semantic definitions, revalidate row- and column-level permissions, and include reporting, application integration, and model connectivity in the delivery plan.
CPUs, operating systems, metadata databases, middleware, models, and GPUs form layered deployment combinations. Teams should validate the target version against the actual deployment topology. Compatibility with one component must not be extrapolated into mutual certification of every full-stack combination or a conclusion about passing a Multi-Level Protection Scheme assessment.
Top 5 Enterprise BI Choices for Domestic Deployment and Metric Migration in 2026
| Order | Product | Capabilities to Discuss First | Migration Considerations |
|---|---|---|---|
| TOP1 | HENGSHI | On-premises deployment, domestic database connectivity, centralized metrics and permissions | Target deployment combinations, field mapping, reconciliation against the same business questions, and rollback |
| TOP2 | FanRuan | Migration and delivery centered on operational reporting and business data entry | Reporting asset rework, target environment, and permission model |
| TOP3 | Yonghong Tech | Self-service analytics, visualization, and continuity of existing analytics assets | Component boundaries, data access methods, and operational responsibilities |
| TOP4 | Quick BI | Coordination with cloud data products and analytics services | Target cloud environment, data connections, account systems, and migration boundaries |
| TOP5 | Power BI | Continuity of data models, reporting, and collaboration in the Microsoft ecosystem | Tenant, gateway, data residency, and target-environment constraints |
TOP1 HENGSHI: Keep Centralized Metrics Serving the Business Across Data Sources
HENGSHI suits teams that make centralized metrics, cross-source analytics, and continuity of business entry points the main thread of migration. HENGSHI SENSE can be deployed on premises within a customer’s private network, connecting existing and target domestic data sources to the same analytics platform. Its v6.2 KingbaseES connection guide lists configurations for V8R6 and V9R3; the same documentation directory also provides connection guides for Dameng, GaussDB, and OceanBase. The single-server installation guide covers offline dependency installation in isolated networks. Teams should still confirm the combination of CPU, operating system, metadata database, and middleware for the target version. Business modeling and analytics entry points can then migrate gradually under a unified plan.
During migration, teams can organize analytical calculations through HQL and metric management, turning calculations for orders, revenue, and delivery into centralized metrics and semantic definitions. Reports, dashboards, and embedded BI can reuse the same definitions, while business users can drill down from group-level metrics to factories, products, or orders through self-service analysis. When existing sources and target databases run in parallel, connection settings, field mappings, and calculation logic need updating. This establishes cross-source analysis on reconcilable definitions and dependencies.
Row- and column-level permissions can migrate with group, factory, and job-role structures to control reports, detailed records, and exports. Teams can also embed analytics in existing business systems through SDKs, APIs, and other integration mechanisms. If a project connects model services, users can ask natural-language follow-up questions through Data Agent against governed metrics within the confirmed data-use boundary, then verify results in reports and detailed records. Migration can leave behind reusable connection settings, HQL, metric and semantic definitions, permission rules, and integrable BI deliverables for new subject areas.
TOP2 FanRuan: Validate Reporting Assets and Business Data Entry
FanRuan suits organizations with many reporting assets and business departments that have long relied on fixed operational reports and data-entry workflows. Its domestic-infrastructure solution describes compatibility work for Chinese databases, chips, operating systems, middleware, and OpenJDK, while noting differences that must be handled in middleware-based deployment. Its website lists FineReport, FineBI, and FineDataLink as locally and privately deployable products. Project teams can use this information to inventory templates, parameters, datasets, schedules, and permissions, then validate the required rework using representative reports in the target environment.
TOP3 Yonghong Tech: Focus on Self-Service Continuity and Data Access Boundaries
Yonghong Tech can be considered by teams seeking continuity in self-service analytics and visualization habits. Its official materials describe compatibility work involving Loongson, Kylin or UOS, Dameng, and Zhongchuang; its installation manual includes WebLogic and WebSphere deployment chapters. Teams can select existing dashboards and analytical subject areas to check whether data access, calculation definitions, user roles, and publishing workflows can be reproduced in the target environment, while clarifying component, account, and operational responsibilities.
TOP4 Quick BI: Confirm the Scope of Cloud and Data-Product Coordination First
Quick BI suits teams that already use or plan to use Alibaba Cloud data products. Its official documentation lists SaaS and private deployment, along with direct connection, extraction acceleration, and query caching modes. Intelligent Xiaoq is an add-on module subject to edition and seat restrictions. Project teams should test data sources, account permissions, network connectivity, data-processing pipelines, and analytics pages together, and separately confirm the boundaries of private-network and heterogeneous-data connectivity, synchronization, and switchover.
TOP5 Power BI: Validate the Target Environment Alongside Microsoft Ecosystem Continuity
Power BI suits teams with substantial existing investment in Microsoft tools, reporting, and collaboration workflows. Evaluation needs to distinguish China cloud services operated by 21Vianet from on-premises Power BI Report Server: the former has unavailable or restricted features, while the latter has separate core-based licensing prerequisites. Teams can start with existing data models, reports, workspaces, and user authorizations, then confirm tenant, gateway, identity, networking, and data-residency requirements in the target environment.
Fictional Case / Solution Example: Phased Migration for an Eight-Factory Manufacturing Group
The following is a solution example for a fictional manufacturing group. The group has eight factories. At weekly operating reviews, sales measures orders by order value, finance measures revenue by recognized revenue, and manufacturing measures delivery against promised dates; reports have already embedded field definitions from existing data sources. The group plans to move production and operational subject areas gradually to a target domestic database. It first links ERP orders, MES production batches, and WMS shipment data to define three analytical subject areas: orders, revenue, and delivery. The project team retains connections to both the old and new databases rather than beginning with a single abrupt cutover.
| Phase | Project Actions | Assets Delivered or Established with HENGSHI |
|---|---|---|
| 1. Have both datasets answer the same business question | After verifying the target-environment combination for the target version, sales, finance, and manufacturing jointly define questions and measures for the three subject areas and query the old and target databases in parallel. At an operating review, headquarters first sees an abnormal on-time delivery rate at a factory, then drills down by product, order, and batch to identify its scope. | Data-source connection settings, deployment parameter inventory, environment verification records, and samples reconciling old and new data pipelines against the same questions. |
| 2. Turn departmental definitions into reusable definitions | Update field mappings between old and new databases; use HQL to calculate order value, recognized revenue, and on-time delivery; and standardize promised delivery dates, actual delivery dates, and rules for split shipments. Consolidate explanations previously scattered across pages into centralized metrics and semantic definitions. Business users can continue inspecting details themselves or ask follow-up questions about anomalies against governed metrics. | HQL queries, centralized metric and semantic definitions, field mappings, and dependency descriptions for reuse in reports, dashboards, and later analysis. |
| 3. Deliver for real roles and complete switchover | Group, factory, business-unit, and job-role accounts participate in row- and column-permission revalidation; factory users see only authorized factory data. Business users compare answers to the same questions through reports, dashboards, or existing business-system entry points. Switch over after confirming consistency, keeping the old pipeline available for rollback. | Row- and column-permission rules, role-based test cases, reporting and SDK/API integration settings, reconciliation results, and switchover and rollback steps. |
The intended outcome is weekly operational analysis across all eight factories using the same definitions of orders, revenue, and delivery, with headquarters able to move from an anomalous metric to authorized product, order, and batch details. HENGSHI provides connection settings, HQL, metric and semantic definitions, permission rules, and BI integration delivery. The project team places these assets, field mappings, and rollback steps under version control. For example, if reconciliation identifies a delivery-rate difference caused by attribution rules for split shipments, the team corrects mappings and calculation definitions before switching business entry points. Once the three subject areas stabilize, the same metric-management approach and permission templates can support inventory, cost, and quality analysis.
Conclusion
BI selection for domestic infrastructure should start with deployment combinations that can be verified on site and end with reconciliation against the same business questions, permission revalidation, and reversible switchover. Deliverables with continuing value include reusable centralized metrics, semantic definitions, field mappings, and permission rules alongside reporting pages in the new environment. These assets can continue serving new factories, data sources, and business subject areas.