A non‐Markovian stochastic Petri net‐based approach to performance evaluation of ontology‐based service composition

Abstract

The Ontology Web Language (OWL)-based web service ontology is one of the major standards for modeling and description of distributed service composition. To analyze the performance of composite service, processes specified in OWL-S give way to tell whether the process meets the performance requirements and to choose the process with higher efficiency from those with similar function. In this paper, we propose a stochastic Petri net-based approach for performance analysis of OWL-S processes, which employs non-Markovian stochastic Petri nets as the intermediate representation. The main innovation of this research includes a translation from OWL-S to non-Markovian stochastic Petri nets and a performance (using expected-process-normal-completion-time as the metric) analysis method. We also validate the correctness of the approach in the case study by showing 90% CI obtained from experimental results that cover corresponding theoretical prediction values.

A time series and reduction based model for QoS prediction of service ontologies

“…23-25, we show the effectiveness of our pre diction framework and make its comparison in terms of prediction accuracy with the traditional approaches [3] [6]. The method for obtaining general distributions of activi ties in [6] is straightforward with steps clearly introduced. While the deterministic distributions in [3] are obtained through averaging historical QoS samples given in Figs.…”

A time series and reduction‐based model for modeling and QoS prediction of service compositions

“…For example, Hwang et al assumed that tasks in composite services follow discrete general distributions, where distributional functions of task execution time map outcomes to values of a countable set. Our prior study assumes tasks that in WS‐CDL flows follow continuous general distribution and adopts a stochastic‐reduction method to predict QoS. Bruneo et al presented a Petri‐net‐based approach for performance prediction and assumed continuous general distribution of task execution time.…”

“…In Figures – , we show the effectiveness of our prediction framework and make its comparison in terms of prediction accuracy with the traditional approaches . These results are obtained through a Matlab 7.0 procedure implementing the Box–Jenkins method on a PC with INTEL Core i5‐750S 2.4 G CPU and 4 G RAM.…”

“…These results are obtained through a Matlab 7.0 procedure implementing the Box–Jenkins method on a PC with INTEL Core i5‐750S 2.4 G CPU and 4 G RAM. Each run of the Matlab 7.0 procedure takes 3.5 to 6 s. The method for obtaining general distributions of activities in is straightforward with steps clearly introduced. While the deterministic distributions in are obtained through averaging historical QoS samples given in Figures – .…”

Fast approaches for semantic service composition in big data environment

“…On the other hand, service compositions in a big data environment face some special challenges. For example, composing stand-alone services across multiple servers can be extraordinarily time-consuming for big data applications (Xia et al, 2012). Currently available approaches in the realm of web and cloud computing, however, cannot provide strict service guarantees for time-sensitive tasks.…”