PIPPIN
The PIPPIN Project
ESPRIT PROJECT 20506
PIPPIN
PILOT IMPLEMENTATION OF PROCESS PLANT INFORMATION WAREHOUSE Lifecycle functional data conforming to STEP AP - 221 Start Jan 96 End Dec 97-Jun 98
17 manyears
Project Partners & Roles
PIPPIN
ICI (UK)- Project Management & Definition of Requirements BP (UK ) - Dissemination Brown & Root (UK)- Deployment EuroSTEP (D) - Architecture, Design & Implementation of Data Warehouse and Application Program Interface Framatome (Fr)( Ingevision - EDS ) - Application Interface ( PDL ) and deployment I+C+S (UK)- Application Interface ( RaPID ) Quillion (UK) - Second Warehouse Implementation Shell (NL) - Standards & liaison
Business Objectives
PIPPIN
Demonstrate & measure benefits from improved utilisation of information Stimulate the market for standards based applications
Technical Objectives
PIPPIN
Design & implement open architecture including :database repository Application Program Interfaces Demonstrate the useability Evaluate the usefulness of STEP Contribute to future direction setting
In short
PIPPIN
A data warehouse that will store any data according to any data model defined in any data definition language and support any view of that data High level API ( business objects, class libraries )
Implemented in both relational ( 6 tables ) and O-O databases with Part 21 transfer between them Potential to use CORBA ( VEGA?) to implement as distributed hetereogeneous database
Requirements
PIPPIN
Data sharing Data exchange Data Integration Lifecycle Specific / class / typical Security / authorisation / authentication Application Program Interfaces
Deliverables
PIPPIN
Requirements specification System Architecture Working implementation Use on trial projects Seminars & Workshops Exploitation plans
Lessons learnt - 1 the bad news
PIPPIN
•AP 221 CD late and poor qualitybecause of poor quality of SC4 methods, process, IR’s •Mapping methods and tools immature •STEP developed ‘bottom up’ not ‘top down’ •Lack of robust, integrated STEP CASE tools •The data and data models of current applications are so awful that it may not be worth developing interfaces - it’s cheaper to rewrite the application
Lessons learnt - 2 the good news
PIPPIN
•Technical basis exists •Pilot implementations work •Several products emerging •Growing industrial demand •Immediate business benefits plus •Large potential benefits
Opportunities for ESCN - 1
PIPPIN
Education & Training on EPISTLE framework & core model AP 221 ARM POSEC/CAESAR and STEPLIB class libraries PIPPIN Implementation architecture
Preparation of additional material video CD Rom - canned demo
Opportunities for ESCN - 2
PIPPIN
Additional R&D PIPPIN API - Basis for data sharing standard Methods of conformance Mapping methodologies and languages Performance improvement
Participate in exploitation EPISTLE working groups - data modelling, standard data, vendor implementations Join PISTEP, USPI/NL Move technology to other sectors
Necessary Improvements
PIPPIN
New SC4 architecture SC4 Quality process Conceptual models and class libraries Mapping methods and languages Universal identifiers and consolidation Improved STEP CASE tools
Strategic Aspects
PIPPIN
‘Trad STEP’ will never do data sharing and integration: and is too big, expensive and poor quality to continue expanding the scope for data exchange We need a manageable number of compatible internationally accepted standards that define the meaning of all information Savings are large ( 20% ), only for early adopters, but information management standards not seen as core business by much of industry
Information Warehouses
PIPPIN
Available products Quillion - PETS Intergraph - NOTIA PrismTech All new, seem very similar, based on EPISTLE methods and either AP 221 ARM or POSC/CAESAR data model. Similar browser capabilities. Performance not widely known - no comparative benchmarks