以過濾及評選模型(FLQR)提升網路服務之品質評估效能

Using the Filter Large Qualify Rest (FLQR) Model to

Improve Quality Evaluation Performance of Web Services

戚玉樑 黃祟德

中原大學資訊管理研究所

桃園縣中壢市(32023)中北路200號

maxchi@mis.cycu.edu.tw

andy_huang@kpec.com.tw

Yu-Liang Chi Chung-De Huang

Dept. of Management Information System, Chung Yuan Christian University

200, Chung Pei Rd, Chung Li, Taiwan, R.O.C., 32023.

maxchi@mis.cycu.edu.tw

andy_huang@kpec.com.tw

摘要

網路服務技術(Web Services)是以文件為中心的應用系統整合方法，它利用訊息的封裝及傳遞，完成異質環境所需克服的互通性問題，因此已逐漸應用在電子商務中；另一方面，多元的服務提供者突顯了服務品質不確定的問題，因此如何濾選合適的服務，則是未來服務仲介者的重要課題。本研究著眼於後網路服務時代的「仲介」問題，提出兩階段的「先過濾再評選」(FLQR, Filter Large Qualify Rest)之改善模型：過濾機制是以減少候選服務之數量為標的，因此考量之因素如服務狀態、穩定性、價格、安全等非功能性因素；其次，評選機制是以服務本身的功能性因素之偏好來決定，並以加權法計算出推薦的序列。由實際的研究結果顯示，使用者透過服務品質的評選機制，將依照品質模型推薦服務的排序，因此可有效解決品質的不確定性。由於一般組織機構皆可建置其區域的服務仲介者，因此本研究的FLQR模型將特別適用於此類服務仲介者的改良。

關鍵詞：服務發現模式、服務品質評估、服務篩選。

Abstract

Web Services architecture (WSA) is an emerging technology for integrating Web-based applications. Due to services are all outsourcings, requesters may have a variety of choices to satisfy their criteria. That is, amount of candidate services advantage the flexibility in implementation. However, requesters encounter the decision issue of selecting a proper service. Motivated by the brokering problems, this study proposes a Filter Large Qualify Rest (FLQR) model to improve a discovery process within a service broker. The model is consists of three components: i). Filter Large is related to remove improper services such as unavailable, expensive, and inefficiency. ii). Qualify Rest is giving a quality evaluation model to compute the performance of each service. iii). Finally, a feedback mechanism offers learning technique to complement FLQR model. In the end, this study utilizes a private UDDI registry center as a services broker. Empirical result suggests that the brokering process in WSA can be improved by using the FLQR model.

Keywords: Service Discovery Model, Service Quality Evaluation, Service Filter