Y. Ishikawa et al. (Eds.): APWeb 2008 Workshops, LNCS 4977, pp. 144 –154, 2008. © Springer-Verlag Berlin Heidelberg 2008 An Approach to Monitor Scenario-Based Temporal Properties in Web Service Compositions* Pengcheng Zhang 1 , Bixin Li 1, ** , Henry Muccini 2 , and Mingjie Sun 1 1 School of Computer Science and Engineering Southeast University, Nanjing, China 2 Dipartimento di Informatica, University of L’Aquila Via Vetoio, L’Aquila, Italy {pchzhang,bx.li}@seu.edu.cn, muccini@di.univaq.it Abstract. Keeping composite services satisfying desired properties has been widely accepted as an important and challenging problem due to the dynami- cally evolving attribute of web service compositions. Runtime monitoring and dynamic verification techniques become first class activities to be performed during the execution of web service compositions. Scenario-based temporal properties depicting the complex interactions among the different services are a kind of important property which needs to be monitored at runtime. However, some complex scenario-based temporal properties cannot be easily represented by the traditional formalism such as temporal logic. In this paper, we first pro- pose to represent the scenario-based temporal properties of web service compo- sitions by the use of a novel notation (Property Sequence Chart). Then, we use Aspect-Oriented Programming techniques to extend the open-source BPEL en- gine (ActiveBPEL) and monitor its execution. Based on these assumptions, we propose a more intuitive approach to monitor the scenario-based temporal prop- erties in web service compositions. Keywords: Web service, Web service compositions, Runtime monitoring, Property Sequence Chart. 1 Introduction One key attribute of web service is that it can support rapid, low-cost and easy com- position of distributed applications even in heterogeneous environments [1]. How- ever, keeping composite services satisfying desired properties has been widely accepted as an important and challenging problem due to the dynamically evolving * The work is supported partially by the National Natural Science Foundation of China under Grant No. 60773105, partially by the Natural Science Foundation of Jiangsu Province of China under Grant No.BK2007513, partially by National High Technology Research and Development Program under Grant No. 2007AA01Z141 and partially by the Program for New Century Excellent Talents in University under Grant No.NCET-06-0466. ** Corresponding author.