Success Case | Semiconductor Factory Data Mid-End, Building Flexible and Efficient Data Service Capability
An integrated circuit manufacturing company with excellent digitalization levels in factory operations and well-established systems. To address data silos and prevent redundant data construction, the company needed to build a unified enterprise-level data warehouse, allowing all data requirements to be extracted from a unified source, saving human and storage resources, and facilitating later maintenance through a single platform for managing data metrics and data quality.
Based on customer needs, Getech real-time collected the company's MES data into the big data platform, enabling high-speed data querying and analysis. This not only helped the customer aggregate data from various business systems but also allowed users to utilize the platform products themselves, integrating data through both real-time and offline methods, and performing data development, consolidation, presentation, and extension within the big data platform.
Previously, the company used Sqoop offline technology to extract MES system data for computation and front-end display, resulting in severe data delays. Through Getech's big data platform, real-time technology enables millisecond-level perception of data changes, second-level collection and computation, and real-time display of production line conditions to workshop personnel, with significant advantages:
2. High concurrency: Task scheduling supports simultaneous concurrent execution of ≥4,000 tasks
3. Large storage: Supports data storage of ≥10PB
Through Getech's big data platform, the enterprise achieved real-time synchronization between the big data platform and MES data, enabling real-time data viewing and analysis without affecting business operations. Meanwhile, one-stop management of data service development, subscription, distribution, and operations broke down business data barriers, improved data consistency, and strongly supported the valorization of data assets, with significant benefits:
1. Achieved zero-delay real-time integration of data from nearly 100 production manufacturing单体 application systems, such as YMS, AMS, RMS, and SPC, into the data lake;
2. Smoothly migrated hundreds of historical data scheduling tasks and scripts to the big data platform, enabling data governance;
3. Based on the data lake, conducted industrial data governance and data analysis applications, starting with virtual metrology and root cause analysis, addressing data development/analysis/governance for 12-inch semiconductor fabs, thereby improving quality and yield.
