Research And Implementation Of Data Distribution Technologies Based On Publish /Subscribe In Distributed Environment

Cyber-physical systems(CPS)is a new intelligent system,which integrates decision-making unit,information-sensing unit and execution unit through complex heterogeneous network.It has data acquisition and transmission,information decision processing,remote behavior control and other functions,can provide us with real-time,safe,reliable and efficient services.CPS network is a data-centric network,data collection,storage,distribution,processing is an important work of the network,and the frequent interaction of information determines the CPS network data with massive,polymorphic,real-time characteristics,fast and efficient data distribution is the application of CPS network requirements,and the existing data distribution model is often difficult to achieve rapid and accurate distribution of data requirements.In view of this situation,this thesis proposes a data distribution method to meet the needs of CPS network,which can improve the quality and efficiency of data distribution as much as the load capacity and computing power of the network are limited.Based on the characteristics of CPS network,this thesis introduces the data distribution service(DDS)as the basic tool of CPS network data distribution on the basis of analyzing the advantages and disadvantages of the existing data distribution model.In this thesis,the DDS is improved from two aspects.In the centralized data distribution service,the paper proposes a hierarchical QoS model and a QoS scheduling strategy,and studies the sub-priority distribution of the data in the distributed data distribution service,the thesis puts forward the method of subscribing data pre-distribution to alleviate the pressure of data distribution according to the structural characteristics of the system.The main research contents are as follows:(1)In the centralized structure publish / subscription system,the thesis proposes a hierarchical QoS model,which divides the QoS requirements of the system into three aspects: data QoS,platform QoS and subscription QoS.On the basis of this,the thesis divides the data into real-time types according to the relevant attributes of the data.Based on the type of strategy and the type of effort,different types of data into different data queues,when there is data need to be distributed,the priority distribution of real-time type Data,and then the distribution of data based on the type of strategy and best-effort type of data,in order to improve the quality of data distribution.(2)In the distributed structure publish / subscription system,in order to alleviate the pressure of data distribution,the thesis puts forward the idea of forecasting the subscription data and then pre-distributing the published data.Firstly,the entropy method is used to calculate the importance and frequency of the items in the subscribed data,which is called the "subscription value".Then,based on the historical data of the data subscription,the time-Data subscription to predict,and finally by analyzing the predicted "subscription value" to guide the pre-distribution of subscription data,in order to reduce the delay in data distribution,improve the efficiency of data distribution.(3)The hierarchical QoS model and the data QoS scheduling strategy are applied to the CPS network data distribution demonstration system.After the system test,the scheduling strategy can achieve the expected effect in the data distribution service.