Research On Multi-Dimention Correlation Modeling For Cloud Computing Environment

Cloud computing has been widely used due to its powerful ability of resource integration and computing.The main research filed in cloud computing environment,including quality of service(QoS)guarantee,security and reliability research,and system optimization management,bring great significance for China.How to maintain the advance for both application and theoretical research,meet the needs of different application areas,and apply key theories and technologies to solve specific problems are problems that needs to be studied more deeply to promote the sustainable and healthy development of the cloud computing.At present,cloud services such as data storage,map navigation,and scientific computing have become more popular.While bringing convenience,a large number of company data,application services,and user privacy are uploaded to the cloud,which also brings serious security problems.The application of cloud security policies and malicious attacks will have an impact on the quality of service of the cloud.Therefore,for users or service providers,it is still needed further research about how to assess the security of cloud systems and the impact of security factors on service quality and how to properly allocate resources to against the impact of security factors on service quality.Second,the application of cloud computing technology in the field of mobile computing has also bring a profound impact.The emerging technology of mobile cloud computing has grown rapidly as the explosive growth of the mobile devices and the demand of the mobile applications.However,the diversity of its application environment and the complexity of its system structure also bring some service quality problems such as service performance degradation and instability,as well as reliability problems such as component failure and network delay.Therefore,in order to ensure the provision of stable,reliable and highperformance services,it is necessary to accurately evaluate the quality of service and establish the relationship between it and the reliability factors according to the characteristics of the mobile cloud computing environment,and finally provide the basis for optimizing resource scheduling and system parameter configuration.In addition,in the research field of reliability,the complexity of the cloud environment makes it difficult and cost to acquire fault data and verificate the reliability evaluation methods in real environments.This situation highlights the necessity and urgency of establishing a reliability simulation experimental platform for cloud environment.Finally,in order to support the development,verification and application of research on quality of service(QoS),security and reliability in cloud environments,the information and operational status of system resources are essential.However,in the existing cloud computing monitoring system,there are problems such as excessive network traffic at some moments due to excessively centralized functions of the master control node,poor system scalability,and inability to cope with node failures in time.In addition,the massive resources existing in the cloud system,as well as the diversity and complexity of the resources themselves,also increase the difficulty of unified management of system resources.Therefore,further research is needed on the cloud monitoring system,the unified identification of cloud resources,and the dynamic joining management of new resources.This paper will focus on the issues raised above.The main research work and innovative achievements of the thesis include:1)Aiming at the problem of lacking security assessment method in cloud environment,a cloud-based association analysis modeling method is proposed,and a securityperformance(S-P)association model in cloud environment is established.First,a modeling approach is proposed for virtual machines,the most important component of the cloud system,to evaluate its security.The model fully reflects the impact of security mechanisms and malicious attacks on virtual machines.Then based on the relationship between virtual machines and clouds,an indicator is proposed for assessing the security of the cloud system.Secondly,a hierarchical modeling method is proposed to establish an S-P correlation model.Queue theory is used to model the performance of cloud computing systems.Then,based on Bayesian theory and correlation analysis,the relationship between security and performance is established,and a new index for evaluating the correlation of complex S-P is proposed.The experimental results verify the correctness of the theoretical model and reveal the dynamics of performance caused by safety factors.2)In view of the lack of suitable performance and reliability assessment methods for emerging mobile cloud computing systems,a correlation analysis modeling method for mobile cloud computing system(MCS)is proposed considering the correlation between performance and reliability(P-R)in mobile cloud environment.Firstly,through the detailed analyzing and understanding of the MCS system structure,the MCS service performance modeling and evaluatingmethod based on multi-queue model is proposed.Secondly,the typical fault factors in the MCS are selected and the reliability model is modeled.Thirdly,based on the Bayesian theory and correlation analysis,the relationship between reliability and performance is established,and a new metric for evaluating the correlation of complex P-R is proposed.The experimental results verify the correctness of the proposed P-R correlation modeling method and reveal the impact of different system parameters on MCS service performance under the reliability factor.Finally,based on the reliability modeling theory,the reliability simulation module is designed and implemented in the CloudSim platform,which provides a general simulation platform for further research on reliability factors in the cloud environment.3)Aiming at the performance improvement of monitoring system in cloud environment and the optimization of resource management,a new cloud resource monitoring system based on Bionic Autonomic Nervous System(BANS),Ontology and multi-valued decision diagram is proposed.First,the Bionic Autonomic Nervous System(BANS)concept was used to design a new cloud resource monitoring system.Policies and mechanisms such as multi-level storage,batch reporting,active discovery,and periodic polling are introduced in the system to reduce system network communication traffic and enable the system to acquire self-organization and self-repair capabilities.Secondly,the Ontology technology is used to describe the various resources of the system and introduce multi-valued decision diagram technology(MDD).The combination of Ontology and MDD gives the system the ability of fast semantic query and the ability of self-diagnosis of complex system faults.It greatly simplifies the monitoring and management of large-scale systems and provides a fast way for intelligent diagnosis of the system.Standardized means.