Deployment Optimization Of Gateway And SDN Controller In Space-Air-Ground Integrated Network

During the past years,space-air-ground integrated network(SAGIN),as an emerging architecture integrating satellite systems,aerial networks and terrestrial communications,has attracted intensive research interest from academia to industry.Thanks to the advantages in terms of large coverage,high throughput and strong resilience,SAGIN can be used in lots of practical fields,such as earth observation and mapping,intelligent transportation system,military mission,to provide wireless access services with high data rates and reliability for amounts of data traffic originated from the increasing new services and applications.However,besides bringing significant benefits for various practical services and applications,SAGIN is also facing many unprecedented challenges like route selection,resource configuration and management,power control,and end-to-end quality of service(QoS)due to its specific characteristics such as heterogeneity,self-organization,and time-variability.Compared to traditional ground or satellite networks,SAGIN is affected by the limited and unbalanced network resources in all three network segments,so that it is difficult to obtain the best performances for traffic delivery.Therefore,the system design and network performance optimization in SAGIN is of great significance.Toward this end,beginning with system design and aiming at network performance optimization,in this thesis,we mainly focus on the important issue of deployment optimization for gateway and software defined network(SDN)controller,from the aspects of data delivery,heterogeneous network management,and resource allocation.It is expected that this thesis could provide theoretical basis and technical support for the integration and development of SAGIN.The main contributions of this thesis are summarized as follows.1.Unlike most existing research works only focusing on the single layer network of satellites,or aerial networks,or the satellite-ground converged systems,in this thesis,we present an integrated architecture based on the three layers of space,air and ground networks.We first discuss the system integration and performance analysis in detail,and then elaborate the characteristics of wireless propagation channel in SAGIN.Toward the gateway selection and controller deployment,some basic models such as bit error rate on satellite link,quality and transmission energy of wireless link,are analyzed.Finally,the applicability SDN technology in SAGIN is explained.2.We study the problems of selecting gateways as relay nodes to optimize the wireless link quality and transmission energy and to guarantee QoS in the communication architecture of cross-layer data delivery from terrestrial network to satellite via aerial network.Considering the traffic amount and distribution,we first formulate the gateway selection of crosslayer communication as a constrained optimization problem under the capacity limitation of space-air link.Then several solutions,i.e.,a basic enumeration algorithm(BEA),a greedy algorithm(GOA),and a particle swarm optimization based algorithm(PSOA)are proposed with the objectives of improving link quality and minimizing link transmission energy.At last,extensive experiments based various settings of link error rate and relative velocity have been conducted and as validated by our numerical results,BEA is able to obtain the optimal results but having extremely long running time,while GOA and PSOA could achieve near-optimal solutions with much lower computational complexity.It can also be seen from the experimental results that,in the cross-layer data delivery architecture of SAGIN,the link error rate,the relative velocity of nodes,traffic amount and distribution,will affect the link quality,the transmission energy and other QoS metrics.3.By introducing SDN,we design a multi-domain SDN architecture for space-air-ground integrated networks to improve the flexibility and programmability of network management,and to enhance the intelligence of SAGIN.Different from most available architectures which usually deployed controllers on the ground in a centralized pattern,our proposed architecture distributes the controllers on the geostationary orbit satellite,high altitude platform and ground network,and divides the whole SAGIN into three SDN domains.The main controller hosted in the ground domain is responsible for handling the central control of the system as a cross-domain orchestrator.Based on the multi-domain SDN architecture,we discuss the design and implementation details and then point out some challenges and open issues.Finally,various performance evaluations are conducted from the configuration updating time on low earth orbit,control protocol packets on low altitude platform,to decision time of gateway selection in cross-layer data delivery.Illustrative results validate that the proposed architecture can significantly reduce the control overhead and improve the network management efficiency.4.We study the joint placement problem of SDN controllers and satellite gateways in the SDN-enabled satellite-ground integrated network(SGIN).It is well known that,in SDN,different controller placement schemes will affect the communication reliability between the control plane and the data plane.In SGIN,information exchange between ground switches,controllers and the satellites must travel through the satellite gateway,and different gateway deployment strategies will produce different propagation delay from the satellite to ground.Thus,the controller placement in SGIN must take into account satellite gateway placement simultaneously.That is,it is a multi-object placement problem in the integrated network,which is significantly different from the single object placement in the ground network.Therefore,the algorithms proposed in previous works cannot be readily adopted.In light of this,we first in this thesis explore the satellite gateway placement problem to obtain the minimum average latency.A simulated annealing based approximate solution(SAA),is developed for this problem,which is able to achieve a near-optimal latency.Based on the analysis of latency,we further investigate a more challenging problem,i.e.,the joint placement of controllers and gateways,for the maximum network reliability while satisfying the latency constraint.A simulated annealing and clustering hybrid algorithm(SACA)is proposed to solve this problem.We construct extensive experiments based on real world on-line network topologies.The numerical results show that,enumeration algorithms could produce optimal results but having very high computational complexity,while SAA and SACA are able to obtain approximate optimal performances with much shorter running time.The simulation results also validate that,in the multi-objective network optimization,it is almost impossible to improve reliability while reducing latency at the same time.The actual application scenarios of the integrated network should be considered when making trade-off between these metrics.