Dynamic Scheduling Framework Of Mimic Service Component In Cloud Environment

With the vigorous development of "New Infrastructure",many important infrastructures are relocated to the cloud.How to ensure the security of applications,resources,and data on the cloud has become crucial.Different from traditional defense methods,the mimic defense principle provides a way to effectively protect the endogenous security of the system.The core idea of the mimic defense principle is a dynamic heterogeneous redundant architecture which is very consistent with cloud resources.Therefore,we consider deployed mimic service components on the commercial clouds with mature service systems.How to efficiently schedule mimic service components in a cloud environment is the focus and difficulty of this thesis.Firstly,this thesis studies the mimic service model system architecture.And introducing the details of the core modules of the system.According to the mimic service model architecture,the microservice architecture is adopted to design the mimic service components.A scheduling management system framework is designed for the mimic service components.Then,this thesis describes the function of each module and the workflow of the dispatch management system module.Secondly,in order to improve the scheduling performance of the scheduling management system,the scheduling distribution module needs to use the prediction results of the offline prediction module,so the prediction accuracy of the offline prediction module plays a decisive role in the scheduling scheme.This thesis uses the combined model of ARIMA model and RBF neural network model to predict the time series of mimic service component resource usage.The strategy-based MCHS scheduling algorithm is proposed,and different scheduling strategies are adopted for special-purpose applications and general-purpose applications.Finally,this thesis constructs a mimic service component scheduling management platform.The correctness and effectiveness of the proposed prediction algorithm suitable for mimic application behavior are verified.At the same time,through the component resource dynamic scheduling experiment,it is verified that the proposed MCHS scheduling algorithm can schedule the mimic service component clusters effectively.