There are at least two challenges with quality management of service-oriented architecture based web service systems: 1) how to link its technical capabilities with customer's needs explicitly to satisfy customers' functional and nonfunctional requirements; and 2) how to determine targets of web service design attributes. Currently, the first issue is not addressed and the second one is dealt with subjectively. Quality Function Deployment (QFD), a quality management system, has found its success in improving quality of complex products although it has not been used for developing web service systems. In this paper, we analyze requirements for web services and their design attributes, and apply the QFD for developing web service systems by linking quality of service requirements to web service design attributes. A new method for technical target setting in QFD, based on an artificial neural network, is also presented. Compared with the conventional methods for technical target setting in QFD, such as benchmarking and the linear regression method, which fail to incorporate nonlinear relationships between design attributes and quality of service requirements, it sets up technical targets consistent with relationships between quality of web service requirements and design attributes, no matter whether they are linear or nonlinear.

Service-Oriented Architecture (SOA) is a loosely-coupled architecture designed to meet business needs of an organization. It is becoming a trend for system development and integration where systems group functionality around business processes. Although SOA does not require Web services, Web services are based on accepted standards and drive SOA to the mainstream. There are at least two challenges with quality management of SOA based Web service systems. One of them is how to link explicitly its technical capabilities with customers’ needs to satisfy customers’ functional and nonfunctional requirements. The second is how to determine targets of Web service technical attributes. The first issue is not addressed at all and the second issue is dealt with subjectively in the current practice of development of SOA based web service systems. Quality Function Deployment (QFD) is a major quality management system used to determine product development characteristics from customer requirements. It has found its success in improving quality of complex products, such as automobiles, aircrafts, and consumer electronics, although it has not been used in the development of SOA based Web service systems. In this paper, we analyze a number of quality of Web service requirements and their related technical attributes, and apply the QFD for developing SOA based Web service systems by linking quality of service requirements to Web service design attributes. An impact based linear regression method is used to determine technical targets of design features of SOA based Web service systems for the satisfaction of quality of service requirements.