Big Data & Analytics for Semiconductor Manufacturing

Transcription

1 Big Data & Analytics for Semiconductor Manufacturing 半 導 体 生 産 におけるビッグデータ 活 用 Ryuichiro Hattori 服 部 隆 一 郎 Intelligent SCM and MFG solution Leader Global CoC (Center of Competence) Electronics team General Business Services IBM

2 Agenda What is Big Data? Big Data in Semiconductor Manufacturing Big Data and Analytics architecture Big Data Analytics use case in IBM Microelectronics Summary

3 What is Big Data? - Big data is about All Data Volume Velocity Variety Veracity* Data at rest Terabytes to exabytes of existing data to process Data in motion Streaming data, milliseconds to seconds to respond Data in many forms Structured, unstructured, text, multimedia Data in doubt Uncertainty due to data inconsistency & incompleteness, ambiguities, latency, deception, model approximations 3

4 Big Data in Semiconductor Manufacturing Fall 2013 Problem statement: Conventional or standard analytical methods and technologies are built for predictive modeling on a small scale, not for investigation of hundreds or thousands of potential factors and interactions Engineers with standard analytical techniques and tools have become the bottleneck, outpaced by data volumes and complexity New methods and software are needed to bridge the gap between analysis and action Automated data mining and analysis tools are needed to explore and uncover problems and opportunities that lead to action and potential manufacturing operation improvements that differentiate one company from its competition

5 Big Data Hadoop There s a belief that if you want big data, you need to go out and buy Hadoop and then you re pretty much set. People shouldn t get ideas about turning off their relational systems and replacing them with Hadoop As we start thinking about big data from the perspective of business needs, we re realizing that Hadoop isn t always the best tool for everything we need to do, and that using the wrong tool can sometimes be painful. Ken Rudin Head of Analytics at Facebook

6 IBM PoV on Big Data and Analytics architecture All Data New/Enhanced Applications Real-time Data Processing & Analytics Operational data zone Landing, Exploration and Archive data zone Deep Analytics data zone EDW and data mart zone What action should I take? Decision management What is happening? Discovery and exploration What did I learn, what s best? Cognitive What could happen? Predictive and modeling Why did it happen? Reporting and analysis Information Integration & Governance Systems Security Storage On premise, Cloud, As a service

7 Transformation to target architecture - start Leverage column-store and in-memory capabilities to improve performance and enable reporting & analysis directly against operational data Data types Actionable insight Operational systems Staging area Enterprise Warehouse Predictive and modeling Transaction and application data Reporting & interactive analysis Reporting and analysis Archive

8 Transformation to target architecture stage1 Provide dedicated processing for faster, deeper analysis and modeling Data types Actionable insight Operational systems Staging area Enterprise Warehouse Deep & modeling Predictive and modeling Transaction and application data Reporting & interactive analysis Reporting and analysis Archive

9 Transformation to target architecture stage2 Leverage Hadoop to capture operational data, leverage additional data types and enable exploration of data prior to normalization Data types Actionable insight Image and video Enterprise content Transaction and application data Operational systems Exploration and landing Trusted data Deep & modeling Reporting & interactive analysis Predictive and modeling Reporting, analysis, content Social data Third-party data Archive Discovery and exploration

10 Transformation to target architecture stage3 Leverage Hadoop for queryable archive Data types Actionable insight Image and video Enterprise content Transaction and application data Operational systems Exploration, landing and archive Trusted data Deep & modeling Reporting & interactive analysis Predictive and modeling Reporting, analysis, content Social data Third-party data Archive Discovery and exploration

11 Transformation to target architecture stage4 Leverage data in motion and streamline processing of extreme volumes Data types Real-time processing & Actionable insight Machine and sensor data Image and video Enterprise content Transaction and application data Operational systems Exploration, landing and archive Trusted data Deep & modeling Reporting & interactive analysis Decision management Predictive and modeling Reporting, analysis, content Social data Third-party data Discovery and exploration

12 Transformation to target architecture stage5 Extend transformation, matching, security and governance capabilities to ALL data Data types Real-time processing & Actionable insight Machine and sensor data Image and video Enterprise content Transaction and application data Operational systems Exploration, landing and archive Trusted data Deep & modeling Reporting & interactive analysis Decision management Predictive and modeling Reporting, analysis, content Social data Third-party data Discovery and exploration Information Integration Metadata & Lineage Information Integration & Governance Data Matching & MDM Security & Privacy Lifecycle Management

13 IBM Big Data & Analytics Offerings Watson Foundations Data types Machine and sensor data Image and video Enterprise content Transaction and application data Social data Third-party data Operational systems DB2, INFORMIX PUREDATA TRANSACTIONS Real-time processing & STREAMS, DATA REPLICATION Exploration, landing and archive BIGINSIGHTS PUREDATA HADOOP Trusted data DB2 WAREHOUSE PUREDATA OPERATIONA L ANALYTICS Deep & modeling PUREDATA ANALYTICS Reporting & interactive analysis DB2 BLU PUREDATA ANALYTICS Information Integration & Governance INFORMATION SERVER, MDM, G2, GUARDIUM, OPTIM Actionable insight Decision management SPSS MODELER GOLD Predictive and modeling SPSS MODELER Reporting, analysis, content COGNOS BI COGNOS TM1 Discovery and exploration DATA EXPLORER SPSS ANALYTIC CATALYST

14 Big Data Analytics approach in IBM Microelectronics Combination of : 1) IBM s Big Data platform and 2) custom applications largely developed, built and driven by IBM Research expertise Leverages all data available in fab: logistics, metrology, inspection, test, tool sensors Equipment Sensor Data Yield analysis routines ~10 Billion data points per day Identifies variables and provides prediction

15 Big Data Analytics use case in IBM Microelectronics Several real use cases are described on following pages Information Warehouse & E-biz interface Demand/Supply Planning Product Demand Management Part Number Build Enterprise Recipe Mgt Equipment Maintenance And Scheduling Energy Management Sensor Systems Manufacturing Execution System (MES) Equipment Control AMHS Control Product Dispatch Engineering Analysis Advanced Process Controls Factory Adaptive Test Engine Process, Measurement and Test Equipment Communications Tools Automated Material Handling Automated Reticle Handling

16 Use Case 1: Big Data approach to the problem of large dataset analysis Traditional Tester Data Ware house Large dataset retrieval Large analysis routine Review reports Challenge: Existing analysis methods struggle with current data volumes pulling and manipulating data takes too long thousands of charts and graphs that require manual review analysis may not be complete before product is shipped New approach In-flight Analytics Tester InfoSphere Streams Near real-time analysis Interactive review Model results in-memory

17 Use Case 1: Real-time multivariate analysis of wafer test patterns with Streams Partial Least Squares (PLS) model compares actual yield to previous results analysis output highlights what has changed Automated Streams solution: Yield Contribution By Pattern compares yield by test pattern to historical data identifies unusual yield behavior, based on multivariate model larger bars indicate larger deviation from historical yield has been used to immediately identify problems on leading edge of new production problem identified before the first wafer had completed testing new data added to existing model and kept in memory for fast and easy analysis Not enough All Goods Too many Partial Goods Benefits: 20% reduction in engineering labor first quality escape prevented - $650k in avoided warranty expense

18 Use Case 2: Adaptive Testing that enables global visibility and decision-making with Big Data From IBM presentation at SemiKorea, Feb2014

19 Use Case 3: Usage of Sensor data in IBM fab for yield control and asset optimization Challenge: Yield learning is the most direct contributor to fab profitability and time to market Huge volume of data (billions of points per day) with many subtle interpretations Want to maximize usefulness of semi-structured tool sensor data for variety of problem solving Large engineering team, with varying skills in analysis, statistics, data mining What we did: Collected and enabled quick review of massive amounts of sensor data, in a simple dashboard Identified tool issues and parameters that influence critical product measurements Developed scoring algorithms, including advanced info theory to highlight relationships ease of use, guides analyst to significant findings Fully automated, with linked reports for full drill-down capability Benefits: Documented savings > $13M during first two years of use Drives actions for tool stability and control, process centering, yield learning, scrap avoidance Systematic implementation has continued throughout the fab

20 Use Case 3: Visualization of Sensor data with scoring algorithms and full drill-down capability

21 Use Case 4: Quality Early Warning System (QEWS) to identify trends in Supply Chain before traditional SPC Challenge Solution Business Value at IBM Quality and supply chain managers need advanced techniques to examine quality date from tens of thousands of parts (incoming, manufactured, deployed) and to provide better, more proactive quality management Software system which uses proprietary IBM technology to detect & prioritize quality problems earlier with fewer false alarms, coupled with push alert functionality for IBM & suppliers to proactively detect & manage quality issues at any stage of product lifecycle Results from QEWS Proof of Concept at external client Cost savings $39M in hard warranty savings, with additional soft savings and benefits in other areas Proactive quality mgt identify and resolve issues before they become problems, up to 6 weeks earlier than traditional SPC Improved quality processes improves quality process efficiency & effectiveness Key Innovations Earlier identification of quality issues through proprietary analytic techniques Fewer false alarms Structured issue prioritization, management, follow-up Distills an ocean of supply chain quality data into prioritized, actionable issues

22 Semiconductor firms see significant opportunities for Big Data to optimize the way they execute GPS across functions Market Research & Product Ideation... align product concepts with consumer desires, improve new product ideas, and new product launch effectiveness for IoT External Data Supply Chain & Distribution... optimize inventory and assets and deliver a reduction in supply chain and distribution costs with single view product Product Development & Manufacturing compress design, development & manufacturing lead time and improve yield and asset utilization Field and Warranty Management... collect field data from connected devices, understand part behavior, predict failures, reduce warranty cost Massive Internal Data Marketing & Sales... design and execute more effective marketing with optimized product assortments, affinities and pricing Procurement & Vendor Management... embed insight into business processes from Manufacturer to Distributor to Customer to Consumer Finance...grow revenue and improve margins with greater business performance insight, and improved forecasting and planning

23 Summary Three Key Imperatives for Big Data & Analytics Success Build a culture that infuses everywhere Invest in a big data & platform Be confident with privacy, security and governance Imagine It. Realize It. Trust It. Focus on business needs Apply how well use data

24 Big Data and Analytics to Cognitive Computing All Data New/Enhanced Applications Real-time Data Processing & Analytics Operational data zone Landing, Exploration and Archive data zone Deep Analytics data zone EDW and data mart zone What action should I take? Decision management What is happening? Discovery and exploration What did I learn, what s best? Cognitive What could happen? Predictive and modeling Why did it happen? Reporting and analysis Information Integration & Governance Systems Security Storage On premise, Cloud, As a service

25