No abstract

No abstract

No abstract

Abstract. Service-oriented computing (SOC) has several unique features which are not typically presented in conventional object-oriented development (OOD) and component-based development (CBD), including the commonalty of service functionality, publish and discovery paradigm of services and dynamic composition of possibly 3 rd party's independent unit services. Hence, OOD and CBD-based analysis and design methods are not effective and expressive enough to model service-oriented applications. Rather, Service-Oriented Analysis and Design (SOAD) has to be defined by using some of the two modeling paradigms and by adding SOC-unique modeling mechanisms. In this paper, we first present a technical comparison among OOD, CBD and SOAD to derive the design criteria for SOAD. And, we define the key artifacts that have to be delivered during SOAD. Based on this, we present a SOAD process which takes service requirements as the initial input and delivers service specifications, compositions, and verified service components as the final deliverables. Each of the five phases in the process is defined with its overview, artifacts, and work instructions. Finally, we present a case study of applying our process to a service domain to reveal its effectiveness and applicability. Once the proposed SOAD is well applied, SOAD artifacts can be more systematically and efficiently implemented with current SOA standards.

Abstract. Component-based software engineering (CBSE) is being accepted as an effective paradigm for building software systems with reusable components. Product line software engineering (PLSE) is an approach that utilizes CBSE principles to support the economic development of several applications in a domain. Hence, the components should conform to relevant domain standards and they must at least provide common functionality of the domain. Moreover, micro-level variability within commonality should also be modeled in components so that a product member-specific business logic or requirement can be supported through component tailoring or customization. Therefore, the degree of commonality and customizability determines the range of component applicability. In this paper, we propose a systematic approach to identify and model commonality and variability (C&V) and present a methodology to reason about the identified C&V model. With the proposed process and guidelines, components in a product line can better support a larger set of family applications.