Research On Load Balancing Of Wireless Multi-hotspot Networks

Through in-depth research on existing systems and existing work,this thesis finds that there are still some shortcomings in the current research of wireless multi-hotspot load balancing sys-*tems based on SDN architecture:excessive use of bandwidth resources in the process of using network bandwidth leads to other users' networks Degraded experience and network perfor-mance;based on existing WLAN network capacity theory and standard models,the different requirements of network capacity and latency have not been fully considered;in the process of users requesting links,the issue of user load balancing and the network are ignored Congestion has even caused the network to be paralyzed;the lack of a pricing mechanism cannot effectively guarantee that each user can obtain a service experience corresponding to their expenditures.To this end,this article combines the SDN framework to start with the protocol and functions,and proposes related algorithm optimization and solutions from multiple aspects.The main contri-butions of this article are:1)Aiming at the research of user behavior characteristics in wireless multi-hotspot net-works,this thesis explores user behavior patterns in multi-hotspot networks.The vast majority still assumes the user's network business and general network topology or channel access mode,while less consideration is given to the unique user behavior patterns in multi-hotspot networks,such as application,AP selection trends,mobility,selfish behavior analysis as well as network experience.This article focuses on the requirements of different users on the network bandwidth and delay in a multi-hotspot network.Based on a deep understanding of user behavior,a fast network switching mechanism(LFHM)based on SDN and load balancing is proposed.In this thesis,simulation experiments are performed through theoretical analysis and a simulation plat-form based on Mininet-WiFi.Through experiments,the handover delay,packet loss rate,and handover stability of different handover methods are compared.The experimental results show that LFHM can effectively reduce the handover delay and packet loss rate between different access points,while improving the stability.2)In order to provide users with better connection and user experience,wireless LAN oper-ators and service providers will always increase the density of wireless access points.Therefore,WLAN users usually find themselves covered by multiple access points and must decide which access point to associate with.Aiming at the problem of network congestion in multi-hotspot networks,based on the existing wireless local area network capacity theory and the 802.11 stan-dard AP access mode,the different requirements of different users in multi-hotspot networks for network speed and network delay are fully considered.Based on the hotspot access algorithm in public multi-hotspot networks,an online association AP selection access algorithm based on game theory is proposed.This article proposes a new hotspot access evaluation model and dis-tributed algorithm,which can optimize the hotspot access through two methods:access control mechanism and network performance analysis and prediction obtained by users after access.At the same time,theoretical analysis and experiments show that the competition ratio of the cor-relation algorithm reaches 1-1/e.Compared with the traditional RSSI-based method,it not only improves the overall throughput,but also solves network congestion,reduces delay,and improves the user's network experience and the effect is obvious.3)Aiming at the problem of multi-hotspot access control,in order to achieve the goals of fair bandwidth allocation and optimized user access rate,the AP comprehensively considers the problem of user load balancing when making decisions on user association requests existing in the network.This thesis proposes a load balancing-oriented access control mechanism based on SDN.Theoretically,a global fair-oriented bandwidth allocation algorithm is used to balance the allocation of resources in the overall planning network between the proportional fairness and the maximum and minimum fairness.In combination with the Bayesian equilibrium theory in game theory,resources are dispatched more reasonably.During the experiment of the algorithm framework scheme,the Mininet simulation platform was used for simulation,and the OpenFlow protocol was extended to enable the AP to send access requests to the controller in time.Cor-respondingly,in order to collect three indicators of signal strength,throughput,and packet loss rate,the intelligent access point association module,AP information acquisition module,and load balancing module were extended on the SDN controller,and multiple connectable accesses were calculated.To determine the best access point to balance the load of each access point and improve network service quality.At the same time,the algorithm runs in the SDN controller,avoiding modification of user equipment connected to the network,improving compatibility and facilitating deployment.In summary,each part of the study is relatively independent and interrelated,and is de-signed to further optimize the deployed network and provide reasonable planning guidance for the network to be deployed.At the same time,it has realistic economic benefits for wireless operators and wireless hotspot network value-added businesses such as shopping centers,cafes,airports,and communities.