Seizing and Sizing SOA Applications with （抓住和上浆SOA应用程序）.pdfVIP

Seizing and sizing SOA applications with COSMIC Function Points Luca Santillo (luca.santillo@) Abstract Service-oriented architecture (SOA) exp resses a p ersp ective of sof tware architecture that def ines the use of loosely coupled sof tware services to support the requirements of business p rocesses and sof tware users. In a SOA environment, resources on a network (e.g. web services) are made available as indep endent services that can be accessed without knowledge of their underly ing p latf orm or p rogramming language. When measuring the f unctional size of a distributed application in a SOA environment, we mustf ace the “boundary po sitioning ” issue. 1st generation FSM methods easily f ail in p roviding a good answer to this question, since they lack a concep tual f ramework p rovided by new generation methods, as COSMI C Full Function Point 2.x and higher. Sof tware applications are not monolithic (any more) . Laye rs, p eer items, and f unctions between them, should be considered, to p rovide a more accurate answer to the sizing question, and hence to the sof tware estimation p roblem. This work aims to describe the boundary p roblem f rom the sof tware measurement p ersp ective, and to p rovide basic guidelines f or the application of the COSMI C Function Point sizing method in a SOA environment. 1. Introduction Functional Size Measurement (FSM) methods aim to provide a technology-independent measure of the size of software systems. Despite (or, because of) such generality, some methods might lack relevant concepts or practices suitable for application in innovative approaches to software development. Service-oriented architecture (SOA) expresses a perspective of software architecture th