Quantitative Assessment of Service Pattern: Framework, Language, and Metrics

Meng Xi,Jianwei Yin,Jintao Chen,Ying Li,Shuiguang Deng
DOI: https://doi.org/10.1109/tsc.2021.3091201
IF: 11.019
2021-01-01
IEEE Transactions on Services Computing
Abstract:For modern service industry (MSI), service pattern is a service provision approach to support the realisation of business model that involves participants from various domains and organizations. A comprehensive description and quantitative assessment of service patterns is of great significance for optimizing the organizational cooperation process in MSI and improving the competitiveness of enterprises. However, most relevant studies on service patterns stay at the level of business processes and qualitative analysis, lacking a comprehensive description of data, resources, and value exchanges among participants. Studies related to pattern assessment focus more on QoS (Quality of Service) rather than consideration of the utility of multi-participant collaboration. Hence, two issues need to be tackled for future development of MSI: a) How to systematically describe and distinguish service patterns with the same business processes. b) How to assess and compare service patterns quantitively and comprehensively. In this article, we propose a service pattern assessment framework which consists of two parts. As part one, we complement the service pattern description language (SPDL) with extended elements and observable attributes to empower it with quantitative analysis, namely Quantitative SPDL (SPDL-Q). In part two, a set of service pattern assessment metrics are designed to assess not only the quality of the services but also the cooperation efficiency of the participants and the orchestration effect of the service patterns elements. The proposed framework was then further validated by a case study, of which four E-commerce service patterns were studied to reveal their evolvement processes. Correlation experiments were also performed to identify the pattern features that have the greatest impact on each metric, so to provide guidance and suggestions for pattern design. Finally, the innovation and significance of the work are outlined and discussed.
computer science, information systems, software engineering
What problem does this paper attempt to address?