支持Web服務合成系統(tǒng)可靠性的工作流恢復機制的研究
全文60頁34206字 敘述詳盡
摘 要
當前，Web服務作為新的分布式計算技術，由于其自身的平臺無關、服務器中立、自描述以及易部署等特性，已經(jīng)被越來越廣泛地接受。許多企業(yè)已將自身的業(yè)務能力以Web服務的形式提供給消費者。在眾多的Web服務出現(xiàn)時，一個新的商機出現(xiàn)了——合成的Web服務，即將現(xiàn)有的Web服務重新整合起來，以形成一個功能更強大的Web服務。Web服務合成是一個企業(yè)通過合成基本的Web服務,來提供給用戶增值服務的能力。
對于合成后的平臺，如何保證系統(tǒng)運行的可靠性是一個必須要解決的問題。Internet環(huán)境的開放式、分布式的特點使得構建于其上的系統(tǒng)更加難以控制；網(wǎng)絡傳輸?shù)目煽啃员旧砭褪艿狡涞讓佑布脚_穩(wěn)定性的制約；Web服務自身異構性、自治性和無狀態(tài)性的特點也大大增加了保證合成平臺可靠性的難度。而且，Web服務合成的實現(xiàn)方法也是影響其可靠性的重要因素。
目前，基于工作流技術的合成已成為Web服務合成的重要方法。工作流是描述一個企業(yè)或組織的業(yè)務處理過程的模型，一直是企業(yè)界和學術界的熱點研究領域。在本系統(tǒng)設計和實現(xiàn)之前，我們在工作流領域已經(jīng)取得了一定的研究成果——完全分布式的面向大規(guī)模復雜應用的工作流管理系統(tǒng)ScopeWork。該系統(tǒng)對包括恢復處理在內(nèi)的多項工作流技術均提出了較為完備的解決方案。
現(xiàn)在，Web服務的合成正是應用了工作流的技術，在完全自治的Web服務上搭建合成平臺。因此，基于Web服務合成系統(tǒng)可靠性的保證完全可以借鑒工作流恢復的思想。本文提出的系統(tǒng)正是基于ScopeWork系統(tǒng)的研究基礎，利用已掌握的、較為成熟的工作流技術來實現(xiàn)Web服務的合成。為了保證合成系統(tǒng)的運行可靠性，我們設計并實現(xiàn)了一個恢復系統(tǒng)，該系統(tǒng)包括檢測模塊、監(jiān)控模塊和恢復模塊，分別完成故障檢測、流程監(jiān)控及恢復處理等功能。為了更好地滿足合成系統(tǒng)可靠性的要求，我們從多個方面對工作流的恢復理論進行了擴展：對系統(tǒng)中可能出現(xiàn)的運行異常重新進行分類；對傳統(tǒng)工作流中的任務結構模型進行了擴展；根據(jù)本系統(tǒng)的特點提出了面向系統(tǒng)不同層次的三層異常檢測模型；針對合成系統(tǒng)中不同類型的流程，我們分別采用不同的恢復策略來處理異常情況。通過這些工作，我們大大提高了系統(tǒng)的錯誤處理能力。

關鍵詞 Web服務合成 工作流 恢復 故障檢測 錯誤處理
Research of Workflow Recovery Mechanism Supporting Reliability of Web Service Composition System
Abstract
Web service, as a new distributed computing technology, has been accepted currently because of its features such as platform independence, server neutrality and self-description. Now many enterprises provide consumers with business ability in manner of Web service. With more and more Web services’ coming forth, a new commerce opportunity, composition of Web services, appears. We can distribute a more powerful Web service, which is composed of existing Web services, to provide added value.
When compositing, we must take into account system reliability. Open and distributed Internet environment makes systems based on it difficult to control; Reliability of network transmission is restricted because of its underlying hardware; The feature of heterogeneity, autonomy and status independence also greatly increases difficulty. Furthermore, the way of composition is also an important factor that will infect system reliability.
Workflow technology has become the most important way of compositing Web services. Workflow, as a hotspot in both academe and business field, is a model that describes handling process inside an enterprise or organization. Before our work on composition system, we have got deeply research in the field of workflow and developed a wholly distributed WFMS ScopeWork, which have put forward a full resolution, including error handling.
Since we use workflow technology to composite Web services, when we try to ensure the reliability of composition system, we can use theory of workflow recovery for reference. The system this thesis advanced is just based on the research of WFMS ScopeWork. To ensure its reliability, we design and implement a recovery system, including detecting module, monitoring module and recovery module, which accomplish the function of failure detecting, process monitoring and recovery handling respectively. To satisfy the reliability requirement, we extend theory of workflow recovery in many aspects, including renewedly classifying runtime exceptions, extending task structure, advancing a three-level failure detecting module oriented different levels of the composition system, using different recovery strategies to handle different types of flows. In summary, we enhance ability of error handling greatly in composition system of Web services.

Key words Web service composition, workflow, recovery, failure detection, error handling目錄
聲 明 I
摘 要 II
ABSTRACT IV
目錄 VI
第一章 引言 1
1.1 問題的提出 1
1.2 當前的研究現(xiàn)狀 4
1.3 本文的組織結構 5
第二章 關鍵技術 7
2.1 Web服務 7
2.1.1 Web服務的定義 7
2.1.2 Web服務的體系結構 8
2.1.3 Web服務的特點 9
2.1.4 Web服務的標準 11
2.2 Web服務合成 13
2.2.1什么是Web服務合成 13
2.2.2 Web服務合成的主要問題 13
2.2.3 Web服務合成的方法 15
2.3 工作流技術 15
2.3.1 工作流的提出 15
2.3.2 工作流管理系統(tǒng) 16
2.3.3 工作流的基本概念 17
2.3.4 工作流管理系統(tǒng)的體系結構 18
2.4 工作流恢復技術 20
2.4.1 工作流恢復的研究現(xiàn)狀 20
2.4.2 先進事務模型（ATM） 21
2.4.3 工作流中的異常及其分類 24
2.4.4 工作流的恢復策略 24
2.4.5 工作流的恢復手段 25
第三章 恢復系統(tǒng)的總體結構 27
3.1 開發(fā)環(huán)境 27
3.2 恢復系統(tǒng)結構框架 29
第四章 日志記錄及流程監(jiān)控 33
4.1 任務結構的擴展 33
4.2 日志倉儲結構 35
4.3 日志的記錄 35
4.4 流程監(jiān)控器的設計 37
第五章 故障檢測機制 39
5.1 “錯誤”建模 39
5.2 三層檢測模型 41
第六章 恢復策略 44
6.1 Web服務合成流程的種類 44
6.2兩種恢復策略 45
總結與展望 48
參考文獻 49
攻研期間論文發(fā)表情況 51
攻研期間科研情況 52
致 謝 53

參考文獻
[TPE01] V.Tosic, B.Pagurek, B.Esfandiari, K.Patel. On the Management of Compositions of Web Services. Network Management and Artificial Intelligence Lab Department of Systems and Computer Engineering, 2001
[CSS01] F.Casati, M.Sayal, M.C.Shan. Developing E-Services for Composing E-Services, HP Labs Palo Alto, 2001
[SGW01] G.Shegalov, M.Gillmann, G.Weikum. XML-enabled Workflow Management for E-Services across Heterogeneous Platforms. University of the Saarland, Department of Computer Science, 2001
[BSh99] B.Benatallah, Q.Z.Sheng, The Self-Serv Environment for Web Services Composition. University of New South Wales, Marlon Dumas Queensland University of Technology, 1999
[PBM02] P.F.Pires, R.F.M.Benevides, M.Mattoso. Webtransact: a framework for specifying and coordinating reliable Web services compositions. Paulo de Figueiredo Pires Rio De Janeiro, Rj –Brasil, April 2002
[CIJ00] F.Casati, S.Ilnicki, L.Jin, V.Krishnamoorthy, M.-C.Shan: eFlow: a Platform for Developing and Managing Composite e-Services. Technical Report, Hewlett Packard Software Technology Laboratory, 2000
[IRL01] V.Issarny, A.Romanovsky, N.Levy. Dependability in the Web Service Architecture. Ferda Tartanoglu, 2001
[MTR01] T.Mikalsen, S.Tai, I.Rouvellou. Transactional Attitudes:Reliable Composition of Autonomous Web Services. IBM T.J. Watson Research Center, Hawthorne, New York, USA, 2001
[BSA00] B.Benatallah, Q.Z.Sheng, H.H.Anne. Declarative Composition and Peer-to-Peer Provisioning of Dynamic Web Services. University of New South Wales Marlon Dumas Queensland University of Technology, 2000
[PBM99] P.F.Pires, M.R.F.Benevides, M.Mattoso. Building Reliable Web Services Compositions. Computer Science Department, COPPE-Federal University of Rio de Janeiro Brazil, 1999
[RSh93] M.Rusinkiewicz, A.Shet. On transactional workflows. Bulletin of the Technical Committee on Data Engineering, 1993
[RSh00] F.Ranno, S.K.Shrivastava. A Review of Distributed Workflow Management Systems. Department of Computing Science, Newcastle University, Newcastle upon Tyne, NE1 7RU, UK, 2000
[SCD02] M.Sayal, F.Casati, U.Dayal, M.C.Shan. Integrating Workflow Management Systems with Business-to-Business Interaction Standards. Software Technology Laboratory HP Laboratories Palo Alto, 2002
[Duw97] W.Du, A. Elmagarmid. Workflow Management: State of the Art vs. State of the Products. Software Technology Laboratory HPL-97-90, July 1997
[AKM95] G.Alonso, M.Kamath, C.Mohan. Failure handling in large scale workflow management systems. Technical report, IBM Almaden Research Center, 1995
[Tyh01] J.Tyhurst. Choosing Transaction Models for Enterprise Applications. Tyhurst Technology Group LLC, 2001
[AKA95] G.Alonso, M.Kamath, D.Agrawal, A.E.Abbadi, C.Mohan. Advanced transaction models in workflow contexts. Technical report, IBM Almaden Research Center, 1995
[Msa87] H.G.Molina, K.Salem. Sagas. Proc. ACM SIGMOD, 1987
[AAA95] G.Alonso, R.D.Agrawal, A.E.Abbadi, C.Mohan. Handling disconnected clients in a workflow management system. In Proc. of the 3rd Int. Conference on Cooperative Information Systems, May 1995
[Ley99] F.Leymann. Supporting business transactions via partial backward recovery in workflow management systems. Datenbanksysteme in Buero, Technik and Wissenschaft, Dresden, Mar. 1999
[SYH00] J.Song, O.Yang, A.S.Hai. An Approach to Optimistic Commit and Transparent Compensation for E-Service Transactions. Vijay Machiraju Software Technology Laboratory HP Laboratories Palo Alto, 2000