サービスアーキテクチャ
DataInsight Industrial Big Data Integration and Analytics Service Architecture Diagram
Starting from the foundational layer, data sources are categorized into two main types: OT data from equipment and IT data from information systems. OT data is collected via edge acquisition tools, leveraging Advantech's Industrial Internet Platform IoT Edge for edge computing and forwarding into the big data platform. IT data is integrated into the big data platform through methods such as JDBC, Restful API, and Web Service. The platform layer enables unified data management.
Once data is ingested into the platform, data integration and computing functions perform filtering, cleansing, and preprocessing. This is followed by data modeling and the construction of various subject-oriented databases around business needs, forming a data sharing platform. The data platform also provides flexible data permission settings to ensure secure data access and supplies data to upper-layer services through multiple access methods.
Against the backdrop of rapid AI development, for AI to truly understand and effectively utilize data, it must be engaged in dialogue using language that aligns closely with business scenarios. This is precisely why an indicator platform is indispensable—it standardizes and centrally manages indicator elements, ensuring that once published, they can be uniformly reused across the entire company. This not only avoids redundant development and saves resources but also fundamentally guarantees data consistency and reliability.
Building upon this, our DataInsight series further develops a rich data application layer. It not only supports various access methods like JDBC/ODBC, API, and MCP Server for convenient third-party system data calls but also provides a pre-built Data Analytics Agent. This agent not only allows users to perform "ask-and-get-answers" style queries using natural language but also conducts in-depth analysis and automated root cause tracing for key metrics. Leveraging Text2MQL technology, the system can automatically translate business questions into precise analysis commands with unified definitions.
This propels enterprises beyond the traditional "data querying" phase towards a complete "analysis-insights-decision-making" closed loop: business personnel and management can independently complete the entire process from data acquisition to root cause analysis, while the data team can focus on accumulating analytical models and knowledge bases. Thereby, it empowers enterprises to make scientific decisions and achieve continuous optimization based on deep insights.