Research On Disaggregated Optical Network Control Mechanism For Data Centers

The emergence of SDN(Software Defined Network)has changed the traditional network control method.However,with the continuous development of cloud computing and big data services,the volume of data continues to increase,and the traffic characteristics are diversified.At this time,network resources need to be dynamically adjusted to ensure the normal operation of different services.Due to the opacity of hardware,the traditional SDN architecture can't achieve the ultimate goal of dynamic controlling network resources.Therefore,a SDN-based new network architecture,DON(Disaggregated Optical Network)has gradually become a research hotspot.It achieves network decentralization by white-boxing network devices,and separates the control,management,and forwarding of hardware devices to promote interoperability between modules of different manufacturers,and better provide network flexibility and agility.At present,research on DON mainly focuses on WAN scenes.In the future,DON will gradually penetrate into the optical interconnection network in the cloud data center.Therefore,how to introduce DON technology into the data center scenario to optimize application performance in the data center has become a hot research topic in the field of optical networks.This thesis focuses on the unique requirements of the application in the data center,and carries out the model research of DON and the design and development of DON control mechanism for application optimization.Specifically,with the goal of application-oriented optimization,designs the DON white boxed node model and completes the control level development,including the implementation of the YANG-based node model,the device driver that supports the NETCONF command generation,parsing and transceiving functions.Finally,through the construction and experimental verification of the experimental platform,it is proved that the developed control plane can effectively support the operation of the proposed DON control mechanism and can effectively optimize the application performance.In addition,for the large-scale network scenario,this thesis designs a fault location solution for traffic flow level,and implements the core fault location algorithm based on traffic flow analysis.By analyzing the underlying network information and traffic flow information,the algorithm can identify the path information of the application flow and complete the accurate analysis of the traffic flow status,thereby distinguishing the fault type and realizing the fault location of the flow level.Experimental results verify the effectiveness of the algorithm.