Research And Design Of Wifi Solution In The Environment Of High Density

With the explosive growth of smart mobile devices, demand towards wlan gets more and more. Deploying wifi has become usual. However, there are still many problems to be solved with traditional distributed wlan and ordinary AC + AP architectures in high density scenarios, such as shopping malls, hospitals, meeting places and other. Problems reflected in three aspects: 1) lacking the ability of controlling the whole wireless network, such as that traditional wlan architectures can not allocate the whole network resources efficiently and reasonably; 2) they do not provide network running states presenting; 3) poor expansibility, distributed networking and ordinary AP + AC products do not provide effective and strongly usable interfaces whitch can be used for the upper layer application development.This thesis proposes a centralized control wifi architecture whitch is called ECWLAN. AP is responsible for data forwarding and station finding. controller is responsible for management, acquiring and storing full mesh state, and allocating power, and channel resources. SDN controller is responsible for the management of user permission and some link parameters. OpenManagement is the communication protocol between the AP and the controller. When AP receives a management frame whitch is neither beacon or probe request, it sends the frame to controller to handle the frame, and also puts the results fed back to the AP. This architecture provides north interfaces to control the whole network. With the north interface we can config channel, power, SSID, passphrase of AP, and also we can obtain link status, load status of every AP, these statistics are the fundmental of channel allocation and power allocation algorithm.In high density scenarios, the traditional channel allocation method can not do optimization considering from a global point of view, and also data source is limited to be gotten. The spectrum utilization is low using only orthogonal channels to allocate channels, due to the limited number of orthogonal channels in some scenarios. On the basis of ECWLAN, this thesis proposes a channel allocation algorithm based on graph coloring, witch is called WCCA. WCCA allows APs working in a partially overlapping channel. Each AP is a vertex, if there is reduplicative coverage and there are stations in thr area, then there is an edge between two vertices. The interference factor is defined according to the number of overlapping regions sta overlapping region, and the degree of interference between both channel. The sum of interference of stations whitch located in contact area between APs is the weight of the edge. Greedy algorithm is used to obtain the optimal solution. In order to verify the reasonableness of the algorithm. The paper implements wifi simulation model using C++. The simulation model supports AP finding and CSMA/CA, and slso achieve LCCS, WCCA algorithm in the simulation based on the model, after comparison, we find WCCA algorithm can effectively reduce the interference of stations in the overlap regions sta.