Research On Architecture And Key Techniques Of Cloud Computing Oriented Simulation System

Modeling and Simulation(M&S)is becoming a critical approach in analyzing complex systems,hence is drawing more attentions of the academic and industrial areas.Many simulation frameworks,models,components and tools are proposed and implemented.Some of them are common toolkits,and some of them are domain-specified.There is little connectivity among these simulation resources and simulation capacity.Meanwhile,there are more and more simulation application with features such as multi-domain,multi-layer and distribution,and engineers have to pay much more efforts and confront with more risks to construct the simulation architectures.The requirement for new simulation architecture that facilitate speeding up the development and reducing the costs is increasing.Architectures based on web and services provide an effective approach to these issues.However,this kind of architecture is facing the computation and storage bottlenecks.Current development of the cloud computing provides a new tools to solving the above problems.Cloud is composed of a lot of computer clusters with high scalability and flexibility,and offers huge computation and storage abilities theoretically by an on-demand way as well as mature technique frame and commercial mode.It is also a distributed programming framework provides underlying supports for distributed system development.It may reduce the development difficulties and decrease the development costs.This dissertation performs a study on cloud computing oriented simulation architecture.It focuses on the three aspects: the system construction,the system formal semantic description and the simulation service aggregation.The main work and contribution of this dissertation are as follows:Firstly,this dissertation compares the state of the art architectures and analyzes the requirements of reusability and interoperability of the simulation architecture.It introduces the ideal and the method to construct the architecture of the simulation system.It also proposes a simulation architecture from a muti-dimentional angle and discusses its structure and execution process.Secondly,this dissertation constructs a formal semantic model for the cloud computing oriented simulation architecture to describe the precise definition of the system structure and execution process.It analyzes the semantic normalization and properties under a consistent domain.This forms the basis of the system business logics with a finer structure.Finaly,this dissertation implements a services organization method in semantic level based on an improved BOM.This method aggregates services with similar functions into a service category cluster and publishes the semantic information of them as a whole.The services search speed is improved,and the diverse user requirements are satisfied.