一個植基於信任聲譽模型和服務水平協議之可靠的網際網路服務組合系統

Abstract

近年興起的服務導向計算使得企業對企業的合作可以過網際網路服務的方式進 行，此外，植基於流程化的架構上，網際網路服務組合的發展已獲得不小的成果並得以 標準化。一般來說，最完整的網際網路服務管理系統包含三項主要活動：服務發現和選 擇，服務水平協議的談判以及服務組合。首先，它必須尋找到符合需求功能的服務，再 使用各種規則如:服務品質、信任和聲譽等參數去做過濾，僅留下少部分最理想的服務。 服務水平協議則可建立服務提供商對消費者的正式承諾，為了加速服務的撮合，自動化 服務水平協議的協商架構是相當必要的，所以該協議最好可以是一個機器可讀的型式。 此外，良好的網際網路服務組合能夠提供企業競爭力的優勢，同樣重要的是，建構服務 組合時也必須檢視該組合是否會引起任何潛在的麻煩。要在一個動態、複雜和充滿敵意 的環境中，組建網際網路服務不是一個簡單的工作，這是由於網路環境的異質性以及隸 屬不同的管轄權所致，因此目前的理論模型和解決方案不是效率不高就是無法強健地處 理各種惡意行為。此外，也沒有太多相關研究討論如何保證和監控服務組合的品質和可 靠度。 只有當不同的應用程序可以自動地選擇和整合，網際網路服務組合的真正威力才能 得到發揮。然而，由於系統的異質性和服務的動態特性產生了很多有待解決的問題和挑 戰。因此，我們提出了一個為期三年的國科會研究計劃以徹底地研究這個議題。 ● 第一年：發展一套優秀的信任與聲譽評估模型以供網際網路服務組合時做選擇 的依據。 ● 第二年：發展自動化服務水平協議的談判架構以供網際網路服務組合時訂定合 同 ● 第三年：我們將使用工作流模式以及第一年和第二年的成果，設計一套可自我 調適的最佳網際網路服務組合機制。

The emerging service-oriented computing (SOC) promises to enable Business-to-Business (B2B) collaborations by means of web services. Recently, the process-based approach to web service composition has obtained considerable attention, momentum and standardization. In general, the up-to-date web service management system focuses on three main activities: service discovery and selection, service level agreement (SLA) negotiation, and service composition. At first, it attempts to discover services with a required functionality and narrow down the selection group to several services using various criteria such as quality of service (QoS), trust, and/or reputation. Then the SLA establishes a commitment to quality levels required by service consumers for service providers to interact effectively. For speeding up the SLA negotiation, an automated framework is necessary for this purpose. Thus, the agreement needs to be in a machine-readable format that can be interpreted at runtime. Moreover, the excellent web service composition is able to provide a competitive advantage for enterprises to drive business revenues. It is also important to detect whether the composition will potentially cause any undesirable feature troubles. However, the construction of suitable value-added services in a dynamic, complicated and hostile environment is not a trivial task. Due to many singularities in the Internet environment, such as the inherent structural and behavioral heterogeneity of web services, as well as the strict autonomy, current models and solutions are not efficient and robust enough to select and coordinate the satisfactory compositions of web services. Furthermore, how to guarantee, monitor, and enforce quality and reliability of web service composition is also not stressed in recent related research results. The real power of integrated web services will be fully realized only when the diverse applications can be automatically chosen and compiled on the fly. However, due to the heterogeneity and dynamic nature of web services, there are a lot of challenges and issues that need to be addressed. Therefore, we propose a three-year National Science Council (NSC) research plan as follows to study this topic. ● First-Year Plan: a smart trust and reputation model for web service selection should be developed. ● Second-Year Plan: towards automated service level agreement negotiation framework for web services composition. ● Third-Year Plan: we will design an adaptive and optimal web services composition mechanism using workflow patterns with the first-year and second-year research results embedded.