The evolution of service-oriented architecture (SOA) and the demands from industry provide opportunity on one hand, and requirements on the other--to develop fault-tolerant, robust, and resilient configurations of web services within an SOA environment. SOA came into industrial practice in past decades; now there are some modern applications that are related to the currently popular Internet of Things (IoT) solutions, although they have performance issues.
The SOA standard contained the notions of choreography and orchestration at the beginning of its evolution [1]. However, there were numerous attempts to create algorithmic solutions in a distributed environment to resolve the problems related to the dynamic configuration of web services, such as how to discover and find services that fit predefined requirements, how to match services that syntactically and semantically fit each other, and furthermore how to keep the performance issues, such as service-level agreements (SLAs), in hand. Once a configuration of web services is constructed, the operational season (the runtime) raises several other problems like how to handle the required changes in a dynamic environment.
The authors provide an answer and proposal for managing faults during runtime. The core of the proposal is the use of chronicles that can be perceived as frames describing events and reactions. The chronicles depict the stimulus response in temporal logic and have a syntactical format that is similar to frames, event objects of concurrently communicating processes.
Vizcarrondo et al. developed middleware based on OpenESB, which is an open-source enterprise service bus. The middleware can interpret the chronicles and the logical rules represented in an ontology that deals with handling runtime faults.
The authors present a case study on an e-commerce application and execute various patterns of web service configurations in order to carry out performance measurements. The results demonstrate the quality of the proposed solution by proving that it is better along several measured parameters than previous solutions.
The paper is interesting for both researchers and practitioners involved in industrial SOA and web service design, and will help with the study of possible solutions for fault tolerance. Researchers of formal models may investigate the similarities and discrepancies between chronicles and well-known process algebras and temporal logic, which can be applied in a dynamic environment.