A study of an adaptive replication framework for orchestrated composite web services

• Published in: SpringerPlus 2013-10, Vol.2 (1), p.511-511, Article 511
• Main Authors: Mohamed, Marwa F, ElYamany, Hany F, Nassar, Hamed M
• Format: Article
• Language: English
• Subjects: Computer Science; Humanities and Social Sciences; multidisciplinary; Science; Science (multidisciplinary)
• ISSN: 2193-1801; 2193-1801
• DOI: 10.1186/2193-1801-2-511

Replication is considered one of the most important techniques to improve the Quality of Services (QoS) of published Web Services. It has achieved impressive success in managing resource sharing and usage in order to moderate the energy consumed in IT environments. For a robust and successful replication process, attention should be paid to suitable time as well as the constraints and capabilities in which the process runs. The replication process is time-consuming since outsourcing some new replicas into other hosts is lengthy. Furthermore, nowadays, most of the business processes that might be implemented over the Web are composed of multiple Web services working together in two main styles: Orchestration and Choreography. Accomplishing a replication over such business processes is another challenge due to the complexity and flexibility involved. In this paper, we present an adaptive replication framework for regular and orchestrated composite Web services. The suggested framework includes a number of components for detecting unexpected and unhappy events that might occur when consuming the original published web services including failure or overloading. It also includes a specific replication controller to manage the replication process and select the best host that would encapsulate a new replica. In addition, it includes a component for predicting the incoming load in order to decrease the time needed for outsourcing new replicas, enhancing the performance greatly. A simulation environment has been created to measure the performance of the suggested framework. The results indicate that adaptive replication with prediction scenario is the best option for enhancing the performance of the replication process in an online business environment.