一個在雲端運算環境中的語意網路服務組合機制

Abstract

隨著雲端運算(Cloud Computing)環境的迅速發展，網路服務(Web Services)的服務品質保證與可信賴度大為增加，因此會有越來越多樣化的網路服務被佈署在雲端上，而當雲端運算環境後端的資料需要被組合運用時，就可以透過網路服務組合的技術來達成。然而也因為網路將會充斥著各式各樣且大量的網路服務，服務組合變成是一個複雜且難解的問題，本研究將針對雲端運算環境提出一個以語意為基礎的網路服務組合方法，運用規劃圖(Planning Graph)的向後搜尋演算法，來找出多組可行的解決方案，並在使用最少網路服務成本的基準下，推薦出較佳的網路服務組合方案。最後本研究針對所提出的機制方法實作一個系統，藉由模擬證實，本文所提出基於規劃圖之向後搜尋式服務組合演算法在三種網路服務產生模型中，可以比傳統的服務組合方法降低94%網路服務的使用成本。

With the rapid development of cloud computing, the quality and reliability of web service increase greatly. Therefore, there will be more and more diverse web services published in cloud. When the back-end data of cloud environment needs to be composited to apply, it can be achieved by web service composition technology. However, because of a very large number of web services in cloud, the problem of service composition has become increasingly sophisticated and complicated. In this thesis, we proposed a semantic web service composition mechanism based on semantic in cloud environment. It utilizes Planning Graph based on backward search to find multiple feasible solutions, and recommends the best composition solution by the lowest using service cost. Finally, we implement a simulation environment to validate the proposed mechanism. Through the simulation results, the proposed algorithm based on backward planning graph reduced by 94% service cost in three types of service generation model to compare with the traditional service composition algorithm.