CERN – European Organization for Nuclear Research Administrative Support - Internet Development Services CET and the quest for optimal implementation and maintenance efficiency Mikael Angberg, AS-IDS with Data warehousing and J2EE Components
CERN IDS Outline Introduction to CET Maintenance and Implementation issues Quality Assurance Maintainability and Versatility Conclusions Questions
CERN IDS 8.6 km Challenges facing CERN Today Build the worlds largest scientific instrument… fewer staff… Staff Budget With less budget.. CERN IDS for more scientists… with higher expectations...
CERN IDS for more financial managers… with higher expectations... Challenges facing the CET Team Today Build the worlds coolest Financial Decision Support application… few staff… Staff Budget With a small budget.. CERN IDS
CERN IDS CERN Expenditure Tracking
CERN IDS CERN Expenditure Tracking “Financial” Managers Globally Distributed Wide range of user needs High Level System Requirements Maintainability Versatility Quality Critical for CERN
CERN IDS CERN Expenditure Tracking
CERN IDS The “System Idea” Application (Java) Data Warehouse User Interface Financial Data Business Logic
CERN IDS Data Warehouse Quality Data and Data Warehouse Quality o Capturing the “right” data o Validating the data o Managed complexity o Error handling o Code Inspections o Coding Standards o Centralised Data
CERN IDS Data Warehouse Quality Financial Data Sources Data Warehouse Stores ExtractionLoading Purchase Finance Process Transform Centralize data One Place, One Format, No Redundancy Error Handling and Validation
CERN IDS Data Warehouse Quality Managed Complexity Reduced number of DB Objects Single Entry Point Parameterized Extraction SUM..GROUP BY Partitioned Table
CERN IDS Application Quality Structured Development Process o Iterative… o Development o Testing o Deploy Coding Standards Code Inspections o Inspired by Fagan and NASA o Based on Sun’s Java Specification
CERN IDS Quality Framework Benefits Greatly Reduced Database Maintenance Increased production code quality (less bugs) Higher developer productivity Ensured Data Consistency Increased transparency and knowledge transfer within and between development teams Extended data quality assurance through automated validation processes
CERN IDS Maintainability and Versatility Objectives Satisfy a Global user community …by providing powerful analysis and reporting capabilities …consisting of more than 1000 people with different needs (and many requests) With a development team of maximum 5 people, minimum 1 - Maintainability - Versatility
CERN IDS Facts and Dimensions Measures / Facts Payments Commitments. Keys Partitioned Fact Table Time Order Dimension Location Supplier Dimension The Dimensional Model Data Warehouse
CERN IDS Maintainability and Versatility Managed Complexity SUM..GROUP BY Partitioned Table
CERN IDS Aggregates and Query Rewrite Raw Data Table Dimension Tables Query Re-write SELECT SUM(sales) FROM raw_data; Aggregate Tables / Materialized Views SUM..GROUP BY
CERN IDS Scalability and maintained simplicity < Partitioned Index Partitioned Table Partitioning < Partitioned Index Partitioned Table
CERN IDS Design Conclusions DWH The data warehouse structure allows : Flexible – Easy to tune and extend Access to any aggregate level “Simple” design – Automation of recurrent maintenance Maintenance and Versatility perspective : Access to any range of data …transparently to the user Scalable – Facilitates growing with preserved simplicity
CERN IDS ART – the Java Reporting Framework Three Main Components “Increase maintainability, by applying existing design/code and documentation standards and ensure in-house knowledge of the product. “ “Use of the framework should be possible with knowledge of standard languages only such as Java, SQL, XML. “
CERN IDS ART – the Java Reporting Framework Web-based Java components XML Templates Web report Input Objects Input Validators Query Builder Report Generator XML Templates
CERN IDS Key “Success” Factors Design assures data quality Development process integrating quality assurance Robust yet Flexible Data Warehouse Design Simplicity - Low on maintenance Component based (Java) reporting framework
CERN IDS Conclusions Users World-wide, highly reliable and available Financial Decision Support Developers Decrease in maintenance efforts Faster “time-to-market” Increased productivity and quality Development focuses on solving business problems – Stable Infrastructure
CERN IDS Thank You Browse to: or For More Information
CERN IDS Summary of CET Web-Based, multi-lingual 20 GB Data Warehouse GB Raw Data Over 1000 active users ~ 1000 Reports per day ~ 24 / 7 Availability 100% Automated Recurrent Maintenance No Ad Hoc Querying