Research On SDN Architecture Design And Key Technologies Of Underwater Sensor Networks

Underwater Sensor Networks(UWSNs)are a kind of systems composed of sensor nodes with acoustic communication and computing capabilities,and is an important technical means for marine data monitoring and processing.Compared with common underwater communication systems,UWSNs have advantages of simple infrastructure,small equipment size,and low cost.Therefore,they have broad application prospects in marine data acquisition,environmental monitoring,submarine resource survey,earthquake and tsunami monitoring,auxiliary navigation,underwater robots,AUV control,etc.UWSNs rely heavily on the equipment hardware infrastructures in common,which have limited services and applications.It is hard to build a multiservice platform,and difficult to manage and maintain.That is,the system risks cannot control after being deployment.As an important measure,redundant deployment is adopted to improve the reliability of UWSNs.However,it creates barriers to the redeployment of UWSNs in the same region and leads to serious resource waste and high deployment cost.Software Defined Networking(SDN)is an innovative network architecture that abstracts network functions and business processes to create a unified and convenient network management based on user programming.It offers a new solution for the development of UWSNs.Based on SDN,an open standard communication protocol and programming interface can be designed,virtualized network reuse can be implemented,and a network architecture in which the control plane and the data plane are separated can be constructed to realize centralized dynamic control of the underwater sensor networks.In the dissertation,we study on SDN architecture design and key technologies of underwater sensor networks.The main contents and innovations list as follows:(1)Design the SDN-based architecture and core device prototyping of UWSNs.It mainly consists of three parts.First,we define the functional requirements and device objects for the SDN data plane and control plane,and then implement network function virtualization based on FlowVisor,and design related SDN communication messages.Second,we implement the hardware of underwater nodes of multi-flow table based on TCAM and construct the software platform based on Open vSwitch.Third,we develop the hardware function module of the surface controllers,realize the software platform of controllers based on OpenDayLight,and plan five core function modules,such as device manager,topology manager,link load monitor,data repeater,and route manager Functional module development.Through on this work,we hope to present the prototype of underwater node and the surface controller based on SDN,which is compatible with the implementation of UWSNs based on user programming.(2)Study an uplink MC-CDMA system of UWSNs and design a multi-user detection algorithm.It includes following parts.First,we construct an uplink MCCDMA system architecture of UWSNs.Then,we design a MC-CDMA transmitter and receiver model.Next,we implement a multi-user detection algorithm of MC-CDMA based on convex optimization,which named COBA.Finally,we carry out simulations to validate the performance of ML,ZF,MMSE,and COBA.The results show that COBA achieves the tradeoff between the detection efficiency and the computational complexity.Based on this work,we hope to build a high-performance communication technology for UWSNs,which provides support for the implementation of SDN data plane.(3)Study a routing scheme supporting for the SDN control plane of UWSNs and present two routing algorithms based on the fuzzy optimization of cut set and path.It includes following sections.Firstly,we build a node-clustering algorithm of SDN named CASDN.Secondly,we implement two routing algorithms of SDN named FCO_MST and FPO_MST based on the triangular fuzzy inference and MST(Minimum Spanning Tree)technology.Thirdly,we compare the performance of the clustering technique and the routing scheme with LEACH,EEHC and Dipanwita's algorithm.The results show that CASDN achieves the tradeoff between the computational complexity and the clustering efficiency.And for the routing algorithms,FCO_MST and FPO_MST are superior to LEACH,EEHC,and Dipanwita's algorithms in terms of the packet delivery rate,end-to-end delay,throughput,etc.Through experiments,we learned that FPO_MST is suitable for UWSNs of sparse deployment,while FCO_MST is more advantageous for the densely pattern.Based on this work,a SDN routing algorithm running on the controller is built to improve the reliability and agility of UWSNs,while reducing the underwater node load,and providing support for constructing the SDN control plane.(4)Study the load-balancing scheme supporting for SDN-based UWSNs of multicontrollers and present a load-distribution technique based on the consistent hashing algorithm.It mainly includes following steps.First,we establish a load-balancing model of UWSNs based on SDN.Then,we construct a clustering coverage algorithm of multi-domains named CCLA&d-CDS by the learning automaton(LA)and the degree constrained connected dominating set(d-CDS).Followed by,we quantify the load type,and then design a load-detection method and a load-distribution algorithm.Finally,we analyze the performance of the load-balancing mechanism and compare the performance of four systems,which are SULA&LB,SUNE,TULA,and TUNE.The results demonstrate that the load-balancing mechanism realizes reasonable and flexible load distribution of multi-controllers.In addition,we know that SULA&LB is an underwater system combining CCLA&d-CDS and the load-balancing technologies,which has advantages in the survival time,bit error rate,the computational complexity,etc.Based on this work,we hope to solve the bottleneck of single controller,and provide support for building SDN-based UWSNs with multi-controllers in large-scales.