Kx FOR SEMICONDUCTORS
Semiconductor manufacturing faces a number of increasing data challenges from managing large volumes of data from multiple sources to generating actions to improve the Fab. Kx technology can handle with ease the demands of fast data in semiconductor manufacturing, specifically the fabrication and wafer test areas.
Our platform is well positioned to provide solutions for Fabs, Process Tools, and Test Areas with functionality for:
• Ingesting increasing volume and velocity of fab data
• Managing disparate data types and sources
• Creating critical KPI’s from complex and numerous data signals
• Enabling real-time visualizations and analytics for decision making
• Characterization of real-time data signals with respect to historical signatures
• Identifying root causes for poor tool performance, process variation, defects and yield loss; especially spatially and temporally
• Actively minimizing yield loss and improving tool/process health using corrective actions
Kx can manage all your data inputs in one platform and enable a complete solution to the fab. In addition, Kx can integrate with your existing architecture to take advantage of high need and value use cases.
PERFORMANCESemiconductor manufacturers are being swamped by the colossal amounts data they have to process, analyze and store as part of the production process. Even more concerning is that it is becoming increasingly more difficult to unlock the rich value of that data as it’s the volume and velocity increases. Kx, with its superior columnar-structured time-series database kdb+ and q, its integrated query and functional language, enables you to scale anywhere from thousands to hundreds of millions of sensors, at any measurement frequency whilst maintaining extremely high levels of performance.
• Ingest and process 20 million streaming sensor readings per second
• Store daily data volumes exceeding 10TB
• Support analysis of trillions of data points
• Process events in sub-millisecond timeframes
WHERE WE OPERATE AND HOW WE ARE DEPLOYEDOur solutions excel in capturing tool trace data, process parametric data, fab yield and test data to enable real-time process monitoring and analytics that improve fault detection, quality and yield. Our solutions also take advantage of the volume and velocity of real-time and streaming data and integrate and augment your Fabs current control and database systems. Kx technology can operate from the individual tool to complete factory level and offer a range of deployment options including upstream, midstream and downstream models for integrating and co-existing with your existing systems
Click here to read more about our flexible deployment options
EXTENDABILITY AND INTEROPERABILITYAn Integrated Development Environment (IDE) and Dashboards enables users to customize and extend product capability for your specific use cases in areas like fault detection, process control and yield ramp where access to large amount of data is critical. A connector framework simplifies integration and coexistence with your existing manufacturing systems, including MES, ERP, and SPC. The solution provides APIs for .NET, Excel, Java, C, C#, C++, Java, Kafka, Matlab, ODBC, Perl, Python, R, Webservices, Websockets, and others to enable you to reuse or augment existing functionality.
ANALYTICSThe Kx platform supports all sensor measurements, tags and attributes to all measurement frequencies at nanosecond precision. A comprehensive library of time-based functions and models support analytics, reports and alerts on streaming and historical data. EmbedPy enables users to use Python and access the latest machine learning models in kdb+. Accompanying HTML Dashboards provide rich visualization and query capabilities on results across multiple devices.
• Capture, analyze and store high frequency time-series data from thousands of sensors to compare with historical data for fault and anomaly detection
• A comprehensive suite of graphical tools with powerful OLAP drill-down capabilities
• Combine historical and real-time data to make decisions from multiple streams