Design And Implementation Of Model Management And Control System In Polymorphic Network

Facing the continuous expansion of Internet application services,the thin waist structure of the traditional TCP/IP network has poor support capabilities in terms of security,manageability and quality of service.In order to solve the existing defects of the IP system,various identifier systems such as geographic location identifier,access identifier,and content identifier have been extensively studied.The Internet has developed "fragmented".At present,it is difficult for a single identification system to provide a universal alternative.Domestic scholars have proposed the concept of polymorphic network,which is based on full-dimensional definable technology to realize the adaptation of network modalities and services,and strive to integrate multiple network advantages.The current related research lacks relevant schemes for network model management and control.Therefore,this paper proposes a design of a polymorphic network model management and control system to achieve comprehensive management of multiple network modalities.Firstly,this paper gives the overall design of the model management and control system based on the overall framework of the polymorphic network and the demand analysis of model management and control.Related module design of the network model management control application and the network model application in the network model management control layer are introduced in detail.The cluster and mapping system design is introduced to enhance the reliability and scalability capability of the system.Then,based on the ONOS controller and the BMv2 software switch,this paper implements each module of the system.Finally,both functional and performance tests of the model management and control are carried out in a physical environment.During the functional test,the results show that the functions of the relevant modules of the system are in line with expectations.In the performance test,the processing latency of the system was firstly tested.The results proved that the average processing latency of the system as a whole was stable at around 3ms.Then a concurrency test was conducted on the system,which proved that the system could still maintain good performance during the 10000 level traffic access.