Federated, Generic Configuration Management for Engineering Data Dr. Rainer Romatka Boeing GPDIS_2013.ppt 1
Presentation Outline I Summary Introduction Configuration Management Overview CM System Requirements Stakeholders Concepts Engineering data Characteristics Processes using Engineering Data GPDIS_2013.ppt 2
Presentation Outline II Proposal: Federated, Generic Configuration Management System Generic Repositories Federation of Repositories Usage scenarios Related efforts Roadmap GPDIS_2013.ppt 3
Summary Situation: Configuration Management of Engineering Data is Not Generic1) Configuration Management is re-invented again and again Different System use different, incompatible approaches Like re-inventing database whenever you need to store data Not Federated1) Silos of Information No Integration between Configuration Management Systems 1) Exceptions prove the rule. Some systems provide some level of generic storage, other systems provide some notion of Federation. Frequently the se terms are used with slightly different meanings in this presentation. GPDIS_2013.ppt 4
Summary - Situation GPDIS_2013.ppt 5
Summary Target: Configuration Management Systems that Store All Types of Engineering Data CM Functionality is Generic Using Application specifies the Type of Data Exchange Data Directly No Exchange, Transfer and Load (ETL) or Manual Process Needed Integrate With Each Other Enables Workflow Integrating Data from Multiple Systems Provide Traceability of Data stored in different Systems Impact of Change to Data in one System on Data stored in Other Systems becomes visible Distributed Web of Engineering Data GPDIS_2013.ppt 6
Summary Proposal: Industry Standard for a Federated, Generic Configuration Management System of Engineering Data Benefits: Compatible CM Systems from different Vendors Integration of Engineering Data Across Enterprise Between Companies Engineering Tool Creators no longer need to re-invent Configuration Management Builds upon existing Standards (e.g. STEP) GPDIS_2013.ppt 7
Summary - Proposal GPDIS_2013.ppt 8
Configuration Management Introduction Concepts and Terminology GPDIS_2013.ppt 9
Configuration Management Overview Configuration management is unique identification, controlled storage, change control, and status reporting of selected intermediate work products, product components, and products during the life of a system Configuration Management Principles and Practice by Anne Mette Jonassen Hass Disciplines using Configuration Management CM : Software Development (Systems) Engineering IT Service Management others Configuration Management System: Software system to enable configuration management Software development: Software Configuration Management (SCM), Source Control System, Version Control System, Engineering: Product Data Management (PDM) or Product Lifecycle Management (PLM) System, IT Services: Automated Systems Administration tool GPDIS_2013.ppt 10
Configuration Management System Requirements Preserve History Previous versions of managed artifacts are preserved Collections of artifacts versions that belong together are preserved Facilitate collaboration Provides information about Who made changes to an artifact, why What are the differences between artifact versions Relationships and dependencies between managed artifacts Determine implications of change Facilitate Data Exchange Allow Transfer of managed artifacts between systems Track relations of managed artifacts in multiple systems GPDIS_2013.ppt 11
Configuration Management Stakeholders Developers of Tools integrating with CM System End Users Using CM System directly Using Tool integrated with CM System Configuration Management Process Developers Configuration Management Control Board Librarian Responsible Manager Configuration Management Administrators GPDIS_2013.ppt 12
Configuration Management Concepts Configuration Item Piece of data stored in CM System, the items managed by the configuration system Synonyms: Artifact, Resource Examples: Part or Instance data (Engineering) Source Code File (Software Development) Word processing or spreadsheet document GPDIS_2013.ppt 13
Configuration Management Concepts Version Series Set of configuration items each of which is obtained by making changes to a previous version (except initial version) Linear Version Series Branched Version Series GPDIS_2013.ppt 14
Configuration Management Concepts Revision Set of configuration items collected for a specific purpose A revision must not contain more than one configured item of any version series Examples: Set of all parts and instances on a specific airplane (BOM) Set of all source code files used for a software release Revision CI GPDIS_2013.ppt 15
Configuration Management Concepts Branch Linear list of revisions each of which is obtained by making changes to a previous revision (except initial revision) Revision Branch CI GPDIS_2013.ppt 16
Configuration Management Concepts Repository System storing configuration items, revisions and branches Revision Branch CI Repository GPDIS_2013.ppt 17
Engineering Data Characteristics Use GPDIS_2013.ppt 18
Engineering Data All data generated during an engineering project Content: Requirements, Engineering Models (for design or analysis), Schematics, CAD Models, Bill of Materials, Contracts, Specifications with Partners, Manufacturing Plans, Project Plans, Documentation, Reports, etc. Software tools to generate data: General purpose: Word, Excel, Visio, MathLab, Mathematica, Special purpose: Modeling tools (SysML, AADL, Modelica, etc.), CAD applications (AutoCAD, CATIA, Creo Parametic, etc.), CFD solvers, Optimization packages, Custom applications: Internal tools developed by/for large companies Pieces of Data are interrelated GPDIS_2013.ppt 19
Engineering Data Example Landing Gear Requirements: The aircraft is capable of landing on runways The forces in the aircraft interior during touch down do not exceed http://1.bp.blogspot.com/_fsvarqsvbd0/s_ijesy-kri/ AAAAAAAABXM/ulihhY_0mlw/s400/747landingGear.jpg Functions: Supports weight of aircraft Allows movement on ground Decelerates aircraft Absorbs shock R1 F1 F2 R2 Configuration Items F3 F4 GPDIS_2013.ppt 20
Engineering Data Example Logical Design: SysML used for Example Other Approaches possible Abstract Models Relates Functions to Parts F1 F2 F3 F4 Configuration Item? http://etd.adm.unipi.it/theses/available/etd-05232006-125203/unrestricted/ Model_Driven_Design_Definition_for_a_Civil_AC_During_Ground_Operation.pdf GPDIS_2013.ppt 21
Engineering Data Example Other Models: Manufacturing CAD Model for Geometry Mesh Representation of Geometry as Input for CFD Support Purchasing CFD Analysis Results: Drag, Noise Generation, Related to Logical Model How to chose Configuration Items? http://www.met.okstate.edu/ FacultyandStaff/Chulho/MEDAL/Pictures/ LandingGear_Chris.png http://1.bp.blogspot.com/-kxhnfpbakv8/ UbopUSdQ8dI/AAAAAAAAAyw/-EbsobedbN4/s320/ aero1.png http://hathor.onera.fr/photos-en/simulations/ images/2011-noise-landing-gear-lagoonelsa.jpg GPDIS_2013.ppt 22
Engineering Data Example Cross-cutting Relationships Example Requirement: Fuel Efficiency (Airplane, not Landing Gear Requirement) Drag produced by Landing Gear (Derived Requirement) Design: Retractable Landing Gear Requirement: Safety: Airplane can Land without Power Safety Analysis Fault Injection Loss of Power Design: http://www.faa.gov/regulations_policies/ handbooks_manuals/aircraft/ amt_airframe_handbook/media/ ama_ch13.pdf Emergency Release of Landing Gear; Gravity Operated GPDIS_2013.ppt 23
Engineering Data Summary Multiple Models of SAME Product Different Purposes Different Fidelity Created by Different Tools Some model complete Product, some only Parts thereof Elements in one Model relate to Elements in other Models Requirement Function Logical Design CAD Design Manufacturing Plan Cross-cutting relations Changes in one Model need to Propagate to other Models Configuration Management System must be able to understand the Structure of the Data in the Models GPDIS_2013.ppt 24
Engineering Data Format Structured Data Each piece of information entered by user has a specific semantic meaning Meaning specified by A language Modeling languages (SysML, AADL, Modelica language, ), Computer language (C, C++, Java, Fortran, ), etc. A schema Database schema, XML schema, etc. A file format Standardized: ISO 10303 (STEP, PLCS, etc.), XMI, SVG, Proprietary: E.g. CATPart, CATProduct, etc. for CATIA Information Model (or Ontology) describes relationships between separately stored data Unstructured Data Word Processing Documents Spreadsheets (structured if content follows rules) GPDIS_2013.ppt 25
Engineering Data Characteristics Large Engineering Project Tens of Millions of Configuration Items Terabytes of Data Tens of Gigabytes of Metadata (attributes, properties) Billions of Relationships between Configured Items Accessed by Tens of Thousands of Users Tens to Hundreds of Thousands Changes per day Trends Designs are getting increasingly complex Number of Relationships will increase More detailed modeling for design and optimization Number of Configuration Items will increase Expect data to grow 10x to 25x over next decade GPDIS_2013.ppt 26
Engineering Data Use Engineering viewed as complex sequence of data transformation processes Safety Analysis Perform Functional Analysis Each Piece of Data can Engineering Model, Input: Requirements Process allocated to Analysis Output: Functions Requirement Output: Safety Analysis Report Create Assembly Generate Mesh for CFD code be a Configuration Item Input: Models of STEP Parts Input: CAD Geometry (e.g. Output: CAD Assembly of file), resolution req. Part-Instances; Notes for Output: Data file with Mesh Manufacturing GPDIS_2013.ppt 27
Engineering Data Use Each Process Manual Automated Input and Output for each Process can be one or several Configuration Items Change of input data may trigger repeat of Process Configuration Management System must be able to understand which Processes use Configuration Items GPDIS_2013.ppt 28
Federated Generic Configuration Management FGCM GPDIS_2013.ppt 29
FGCM Requirements Summary Configuration Management System Repository for Storing Configuration Items, organized in Version Series and referenced by Revisions organized in Branches Provides Access Control, Audit Capabilities, etc. Generic Must NOT impose Structure of Data stored in System Data Structure is Supplied to System Mechanism for Data Structure Specification must be flexible enough to cover all Engineering Data Formats, Languages, Mechanism for Data Structure Specification must be able to express relationships between Configuration Items Federated Provides Mechanism to synchronize Data in different Repositories Supports Partial Synchronization (Some but not all Data) Repository Access through Standards Based Interface Interoperability of Implementations by different Vendors GPDIS_2013.ppt 30
FGCM Data Structure Specification Mechanism FGCM provides Type System Build-in types: Boolean, Integer, Floating Point, String, User-specified: Records, Arrays, Lists, Maps, References: Specify relationships Functions: Represent processes Type System Theory part of Computer Science Primarily used for Programming Languages Can be used to derive guarantees what data structures can be expressed Will re-use existing mechanisms to specify a type system Unified Modeling Language Class Diagrams (UML) XML Schema Definition (XSD) Eclipse Modeling ecore Classes Web Ontology Language (OWL) Every Configuration Item has a FGCM Type GPDIS_2013.ppt 31
FGCM Data Structure Specification Mechanism Data Formats, Languages, Information Models evolve over Time FGCM Types must be able to change Apply Configuration Management to FGCM Types A FGCM is a Configuration Item A System in which every Configuration Item has a FGCM Type and every FGCM Type is a Configuration Item can be defined consistently. When FGCM Type changes Data described by the Type must be updated FGCM Type Change includes Function for Data update Different modes for Data Updates: Immediate, Delayed, GPDIS_2013.ppt 32
FGCM Federation Requirement for Federation 1. Some Use Cases require the Capability to synchronize Data between Repositories after they have been established independently 2. Addresses System Scalability Number of Configured Items: We don t need one Repository capable of holding all data Number of Users: New synchronized Repositories can be added to support more users Federated Configuration Management has proven successful for Software Development GIT is widely adopted (Linux development, GITHub) GIT is a federated source control system GPDIS_2013.ppt 33
FGCM Federated versus Centralized Centralized: Single Repository All Clients access central Repository Repository coordinates Client interaction One Pre-Defined Configuration Federated: Any number of Repositories Every user has a local Repository Synchronization between Repositories provides Coordination Configuration can be adjusted to meet Requirements GPDIS_2013.ppt 34
FGCM - Architecture Integrated Engineering Tool FGCM UI FGCM API FGCM API FGCM Repository API FGCM Repository Reference Implementation Single User Local Repository Store Shared Server Repository Database Store Shared Server Distributed Repository Store FGCM Repository Commercial Implementation FGCM Repository Legacy System Wrapper Implementation Existing CM System FGCM Repository API Data Access / Manipulation: Configuration Items, Revisions, Administration: Types, Permissions, Repository to Repository Synchronization GPDIS_2013.ppt 35
Ad-hoc Collaboration Suitable for casual collaboration / prototyping Example: Bob has a new Design Idea Discusses it with John and Jeff The get permission to explore the Idea They need expertise from a different field, pull in Sandra to run simulations Federation of local Repositories facilitates Data Exchange, Collaboration Supports expanding Team Prototype Data (incl. History) can be synched with Main Repository if the Idea proves successful GPDIS_2013.ppt 36
Central Deployment Federation supports Configuration with single, central Repository Appropriate if One Authorative Source for all Data is needed Server must use scalable Implementation of FGCM API GPDIS_2013.ppt 37
Multi-tiered Deployment Still provides One Authorative Source for all Data Provides better Scalability than single central Repository Number of Users Amount of Data GPDIS_2013.ppt 38
Chained Deployment, Partial Synchronization Still provides One Authorative Source for each Type of Data Not all Logical Design Data is needed for Physical Design Not all Physical Design Data is needed for Manufacturing Planning Still Provides Traceability across the whole Design Provides better Scalability than single central Repository Number of Users Amount of Data Can be combined with Multi-tiered Deployment GPDIS_2013.ppt 39
Data exchange between Organizations Contractor deploys Exchange Repositories for Data exchange with Subcontractor Only Data that this Subcontractor may see is synchronized to the Exchange Repository Repository at Subcontractor is synchronized with Exchange Repository Repositories at different Organizations may be different FGCM Implementation from different Vendors May lead to World Wide Web of Engineering Data GPDIS_2013.ppt 40
Related Efforts Data Exchange Formats Model Checkers Data Unification Über Repositories GPDIS_2013.ppt 41
Related Efforts Data Exchange Formats Examples: STEP, PLCS, XMI, Complementary FGCM will provide Support Formats are well-defined Can be mapped to FGCM Type System FGCM supplemental Standard will provide Binding of Data Exchange Format to FGCM Type(s) Defined in Conjunction with Standards Committee for Data Exchange Format Provides Import / Export capabilties FGCM will support Data Exchange without ETL process GPDIS_2013.ppt 42
Related Efforts Model Checkers Model Checkers can validate that Multiple Models are consistent Internally With each Other Checking of Consistency of Multiple Models can be used as Alternative to Synchronization of Multiple Repositories May require translation of Models in different Modelling Languages May require Meta-Model describing Relationships of lower level Models Choice based on Process Preferences Availability of Tools Model size and Checker Performance GPDIS_2013.ppt 43
Related Efforts Data Unification Meta Language to Translate between similar Models modeling the same or similar Content Example: FUSED Complementary Meta Language and FGCM Type System could be integrated Research Project GPDIS_2013.ppt 44
Related Efforts Über Repositories Single Central Repository collecting Data from other Repositories / Tools Example: SPEEDS, Share-A-space Two Main Variants Support Limited (Standard) Data Formats Data Format Agnostic Competitive Disadvantages: Less Flexible Scalability Single Points of Failure Suitable only for large organizations GPDIS_2013.ppt 45
FGCM Roadmap Organize Interested Parties Now to 1.1.2014 Future Users of FGCM Standards Organisations PDM / PLM Vendors Engineering Tool Vendors Academia Contact rainer.j.romatka@boeing.com Draft FGCM Standard Interfaces Starting 2014, 6 to 12 months Develop Prototype Starting 2nd Half 2014, 2 to 4 years Use Prototype in Pilot Engineering Projects Validate Usability Revise FGCM Standard Formalize FGCM Standard Standards Body Selection Formal Specification Suite of FGCM Standard Validation GPDIS_2013.ppt 46