150 likes | 333 Views
Prof. Björn Regnell Lund University Sony Ericsson. Requirements Engineering in platform subcontracting. General trend in embedded systems industry. The number of lines of code is increasing rapidly; - we can not develop everything ourselves… Non-functional requirements are critical!!.
E N D
Prof. Björn Regnell Lund University Sony Ericsson Requirements Engineering in platform subcontracting
General trend in embedded systems industry • The number of lines of code is increasing rapidly; - we can not develop everything ourselves… • Non-functional requirements are critical!!
Requirements Engineering and Roadmapping • Pre-development and roadmapping in line functions; development in projects Portfolio Design Market Research Product Management Product Planning Application Planning Design Studies Product Design Hardware Dev. Technology Roadmapping Sub-domain RE Application Platform Dev. Software Dev. Technology OutlookTechnology Platform REStandardization TG Define Launch Time Toll Gate Concept TIMING CHALLENGE!!!
Layers of Platforms Product B Rel n Product A Rel m Application Platform (HW+SW) Release j Application Platform (HW+SW) Release j Application Platform (HW+SW) Release j In-house platform Company Y Technical Platform (HW+SW) Release i Technical Platform (HW+SW) Release i Company X Technical Platform (HW+SW) Release i Subcontracted platform
Platforms as product lines => Synchronization Challenge Mobile product projects Internal system & application platform Technical platform product Internal platform (HW/SW code base) Time Milestone
PlatformCompetitor Operator DeviceCompetitor Requirements in sync over borders? Consumers Retailer Integrator Supplier Market Req. Market Req. Roadmaps Roadmaps ArchitectureReguirements Product Management Product Management DetailedReguirements RequriementsSub-domainworking groups RequriementsSub-domainworking groups SoCStatement ofCompliance TechnicalPlatform Development IntegrationPlatformDevelopment SRSSystem Req Specfor integration platform Hotlist SRSSystem Req Specfor technical platform Order Release Change Request Maintenance Release
Too much Dealing with very large requirements databases Requirements Database Strategic? Profitable? Related? Ambiguous? Group? Complete? Split? Reject? Expensive?
NFR challenge • Systematic prioritization of FEATURESis state-of-art in roadmapping and platform/product scoping • …but… • Prioritisation of QUALITIES is handled ad hoc with no specific support for NFR roadmapping
Cost view Benefit view Excessive Competitiveadvantage Saturation breakpoint Quality level Useful barrier Differentiation breakpoint Useless Quality level Utility breakpoint Roadmap view Competitor B Target release n2 Competitor A Target release n1 Bad Target Current Quality level (Feature X, Segment Y) Utility Differentiation Saturation QUPER for Quality Roadmapping
QUPER Expected Gains • Making tacit requirements explicit • Coherent terminology across organizations • More qualified scoping debate • Documented rationale • Support change management • Better decisions -> More profitable products
Future technology needed for: • - Quality requirements • - Natural language requirements • - Dependencies among requirements • - Finding your way in very large requirements databases • - Cross-organizational requirements engineering processes
Further reading • [IST2007] "Requirements Engineering Challenges in Market-Driven Software Development – An Interview Study with Practitioners", Lena Karlsson, Åsa G. Dahlstedt, Björn Regnell, Johan Natt och Dag, Anne Persson, Information and Software Technology, 2007. • [EMSR2005] "Market-Driven Requirements Engineering for Software Products", Björn Regnell and Sjaak Brinkkemper, Chapter in Engineering and Managing Software Requirements, Eds A. Aurum and C. Wohlin, Springer-Verlag, ISBN 3-540-25043-3, 2005. • [REFSQ2007] “A Quality Performance Model for Cost-Benefit Analysis of Non-functional Requirements Applied to the Mobile Handset Domain”, Requirements Engineering: Foundation for Software Quality (REFSQ2007), Lecture Notes in Computer Science, Volume 4542/2007, pp. 277-291. • http://www.cs.lth.se/home/Bjorn_Regnell/->follow link to Publications
Thank you! bjorn.regnell@cs.lth.se bjorn.regnell@sonyericsson.com