The Incompatibilities between Software The Incompatibilities between Software The Incompatibilities between Software Component Based Development and Present Component Based Development and Present Component Based Development and Present UK MoD Procurement Approaches UK MoD Procurement Approaches UK MoD Procurement Approaches Michael Looney Department of Information Systems University of Portsmouth University of Portsmouth
Presentation Presentation n Background Background n Paradigm Shift Paradigm Shift Procurement & Management Issues Procurement & Management Issues Summary Summary University of Portsmouth
Software Problems Software Problems Software Problems Intangible n invisible, l flexible ‘easy to change, l Discontinuous Failure Modes n l single error can cause system failure l 100% correctness required Complex n l many levels l many modules l millions of lines of code Hardware first n l traditional systems started with hardware first l software was expected to take up the slack University of Portsmouth
Traditional Review Process Traditional Review Process Concept System Approval Acceptance Initiation Development Build Field Progress Reviews University of Portsmouth
Lifecycle Timeline Lifecycle Timeline Lifecycle Timeline Operation Disposal R&D /Development Operations and Maintenance 30 % LCC 70 % LCC 25- 50 Year Service Life 5 - 10 Years University of Portsmouth
System Complexity System Complexity 1950-60s 1970-80s 1990-00s Dedicated Subsystems Federated systems Integrated Systems Digital Fire Control Flight Control Aircraft wide information Pt-Pt Wiring Fly by Wire integration Crew dominated operations Crew-assisted operations Massive data bases Digital sensor processing Integrated diagnostics 64KB 1MB 100MB University of Source: US Air Force Research Laboratory Portsmouth
System Comparisons System Comparisons System Processors Memory ADA1(1961) 3 36K ADA2 2 + 3 96K ADAWS 2 + 2 600K CAAIS 1 100K CACS 2 + 7 1M SMCS ~150 ~100M SSCS(1997) ~300 ~400M University of Portsmouth
Presentation Presentation Background Background Paradigm Shift Paradigm Shift n n Procurement & Management Issues Procurement & Management Issues Summary Summary University of Portsmouth
Changes Changes Process n Control over the outcome n System costs n University of Portsmouth
Paradigm Shift Paradigm Shift System developers/procurers have moved from being producers to being consumers University of Portsmouth
Producer v Consumer Producer v Consumer Producer Consumer • Identify requirements • Identify requirements • Build bespoke components • Framework/Integration Strategy • Identify unique interfaces • Adopt standard interfaces • Integrate bespoke components • Procure components based on standards • Field and support bespoke system • Integrate Components into Framework • Field and support integrated system University of Portsmouth
A Necessary New Way of Doing A Necessary New Way of Doing A Necessary New Way of Doing Business Business Business Traditional Required Approach for Development Approach COTS Based Systems System System Context Context Simultaneous Architecture & Definition and Design Tradeoffs Architecture Marketplace & Design Implementation Adapted from Oberndorf & Foreman, SEI, 1999 University of Portsmouth
Systems Integration of Components Systems Integration of Components Systems Integration of Components Real World Systems of Systems University of Portsmouth
Component Based Development Component Based Development New Development Items Legacy Components Commercial Components University of Portsmouth
Architecture/Interface/Glue Code Architecture/Interface/Glue Code Architecture/Interface/Glue Code Fully New Development Wrapped Item Architecture Added/Modified Interface Glue code Legacy component Legacy component Newly developed component University of Portsmouth
Process Model Process Model Process Model Specify Requirements/ Constraints Produce set of feasible System Designs Project Management, Configuration Control Analyse System Negotiate Trade off at Technical System level & Prototype solution Presentation of Alternative Implementation/ Options for Decision Making Acceptance Process at Business level University of Portsmouth
Sub Process Model Sub Process Model Sub Process Model Produce set of feasible System Designs Selection of Arch./Framework Identification of Glue Code Standard Identification of Interface Standards Impact Analysis, Derived Constraint, Human Factors Combinatory issues Socio/economic Implications Component Selection Component Tailoring University of Portsmouth
Presentation Presentation Background Background Paradigm Shift Paradigm Shift Procurement & Management Issues n Procurement & Management Issues n Summary Summary University of Portsmouth
Management Process Management Process Management Process Initiating processes Planning processes Executing Controlling processes processes Closing processes University of Portsmouth
Management Changes Management Changes For Against • More sources • Better quality • Not finding precisely the right component • Newer technology • Market driven changes • Cheaper implementations • Vendor support for the component stops • Faster availability • Easier Interconnections University of Portsmouth
Software Project Managers’ Software Project Managers’ Balancing Act Balancing Act Trade offs between Impact of requirements, high reliance Strategies components and on COTS for through cost components life support Strategies to deal with Components, component architecture, obsolescence integration University of Portsmouth
Procurement Procurement Procurement Acquisition is a whole life process, covering requirements setting, initial procurement, in- service support and disposal University of Portsmouth
Smart Procurement Smart Procurement Smart Procurement Through Capability Acquisition will Teamworking be Whole life Faster approach Cheaper Using best practice Better University of Portsmouth
- Quality + Mythical or Real ? ? Mythical or Real System Requirements Area of Flexibility Better, Cheaper, Faster ? Schedule Cost - + + - + Performance Adapted from an NUWC presentation University of Portsmouth -
Smart Procurement Stages Smart Procurement Stages Smart Procurement Stages Concept Stage : First stage which forms the integrated team and produces the n user requirements. The business case is assembled for Initial Gate approval. Assessment Stage : Begins after Initial Gate, risk is reduced to a level n consistent with delivering an acceptable level of performance to a controlled time and cost. The business case is assembled for Main Gate approval. Demonstration Stage : During this stage the ability to produce an integrated n capability is demonstrated. The prime is selected and a contract based on the system requirements placed. Manufacture Stage : The integrated team deliver the solution to the military n requirement, completing system development and production. System acceptance is conducted. In-Service Stage : The line management provide effective front line support n and carries out approved upgrades or improvements, refits and acquisition increments Disposal Stage : Efficient,effective and safe disposal of the system n University of Portsmouth
Smart Procurement Smart Procurement Smart Procurement Initial Main Contract Gate Gate Let Concept Assessment Demonstration Manufacture In-Service Disposal Time University of Portsmouth
Smart Procurement Smart Procurement Smart Procurement Progressive Acceptance User and System requirements Design Certification System Acceptance In-Service Date University of Portsmouth
Smart Procurement Definitions Smart Procurement Definitions Smart Procurement Definitions n Initial Gate l A relatively low approval hurdle, between Concept and Assessment, intended to encourage early and full exploration of a wide range of options for meeting a particular capability. n Main Gate l An exacting approval hurdle, between Assessment and Demonstration. A business case case at Main Gate should recommend a single technology and procurement option. University of Portsmouth
Maintenance/Upgrade Issues Maintenance/Upgrade Issues Maintenance/Upgrade Issues Environment Platform Programme availability Cost Risk Operational Modification requirement task Capability Obsolescence upgrade change Drives Technology Technology Enables/causes availability availability Dictates Implies Constrains University of Portsmouth
Change Drives/Rates Change Drives/Rates Change Drives/Rates GSAW Survey Release Frequency (months) 1999 6.3 2000 8.5 8.75 2001 1 Adaptive maintenance often biggest CBS life cycle cost 2 Average of 3 releases before becoming unsupported Ron Kohl survey Ground Systems Architecture Workshop: 2002 Aerospace Corp., LA March 2002 University of Portsmouth
Recommend
More recommend