Monitoring And Prediction Of Self-adaptive Software Systems Based On Web Services

Along with the internet penetrating expansively to all corners of society, software development is also faced more challenges with a variety of new application models emerging, for instance the gradually increased in the management and maintenance costs, the gradually increased in the demand for robustness, obviously differences between the state of storage and operation..A self-adaptive systems, especially the web services-based adaptive systems response to these challenges providing a series of new settlement mechanism to make the software itself to collect changing information on real-time and automatically adjust itself to provide better services for users based on pre-set strategy, if necessary.On the basis of previous work, we proposed a monitoring model for A Self-adaptive System Based On WebService, which could be better to meet users’ needs and the change of the operating environment. Meanwhile, on the basis of the monitoring model, we proposed prediction methods of variable Exponential Smoothing for supporting QoS parameters of web services based on the ratio. The main research of this study included following aspects:1、A proposed monitoring framework of a Self-adaptive System Based On WebService (SASBW)To accomplish adaptive of Web services-based adaptive software systems, you must know the changes in demand and system operating conditions and you must design an effective monitoring mechanism and prediction mechanism to obtain the above information. Our presents a workflow-based monitoring framework of adaptive software system.The framework consists of two sub-modules:Web services monitoring and context monitoring, Its main features include:①Web services monitoring module,which is mainly to classify and model the QoS indicators in the implementation of Web services in accordance with its different collection methods; to process according to the different acquisition and method of calculation defined in the QoS model,then to obtain and store the QoS metrics.②The context of environmental monitoring module,it responsed for the monitoring of context in the process execution, in this paper it given the description file of the context,from which we could obtain context information in order to promptly adjust processes to meet the changes in system dynamics.③Prediction module,it is mainly responsible for the definition of the prediction point,data acquisition and processing,providing predictive model for forecasting data.2、A proposed prediction method of dynamic Exponential Smoothing based on the ratioIn our study, we propose an improved prediction model based on exponential smoothing prediction method. This improved prediction model is to consider the improvements from the perspective of data transformation, some of the technical processing of your data can be done first before the use of exponential smoothing. Here,it use computational sequence ratio to form a new sequence with exponential smoothing forecast; Simultaneously, in the forecasting process it dynamiclly generates weighted coefficient of the prediction model in accordance with the the relative error of predicted values and actual values.3、Achievement of monitoring and prediction of Self-adaptive System Based On Web ServiceCombined application instances, we use Eclipse graphical development tool and QoS information collection tools automatically generated from the WSDL, we developed a workflow-based self-adaptive software system, principally realize the Web services monitoring module and context monitoring module as well as the predict module of Web service indicators. Combined with real-life tourism,we can customize the case to further validate the monitoring framework, to predict the feasibility and superiority of the method.