Towards automatic derivation of a product performance model from a UML software product line model

摘要

WOSP provides a forum at the intersection of software and performance to bring together software engineers, developers, performance analysts and modelers and to generate insights into the difficult and urgent problems in this area. Modern applications add new challenges every day, and demand new approaches for modeling and analyzing the performance of design alternatives in all the phases of the software lifecycle. We have to cope with distributed and mobile applications, sometimes with heterogeneous and intermittent connections, increasing system complexity, rapidly evolving software technologies, short time to market, incomplete documentation, and poorly defined requirements. >This is the seventh in the series of International Workshops on Software and Performance that began almost ten years ago in Santa Fe, New Mexico, in 1998. The research agenda identified at that meeting has in fact propelled the research field in the decade since, notably in the area of performance evaluationof designs. But new challenges have emerged, particularly in the provision of flexible services with strong time-delay requirements over the web and over wireless connections. >This year's technical program is as strong as ever. We received 39 submissions which were thoroughly reviewed by three or four reviewers from an expert program committee. From these 15 were accepted as research papers, 2 as experience papers and 3 as short papers. The workshop is structured in 10 sessions over 3 days. Session topics range from Performance Diagnosis and Improvement to Modeling from Component Libraries. A particular strength of this program is its combination of papers on measurement and on modeling. We are confident that this body of knowledge will advance the state of the art in the field and contribute to bridging the gap between research and industrial practice. Only the synergy between these two worlds can provide real progress. >This program echoes the first workshop in providing ample opportunities for open discussion of issues and research directions, prompted by panels. We have identified three areas as key to the future development of the field: (1) identifying new industrial issues, (2) testing, empirical assessments and benchmarks, (3) unifying the use of measurement and models.