Initial Evaluation of Load Balancing using Network Address Translation for Ad-hoc Wireless Positioning Network

We have developed an Ad-hoc Wireless Positioning Network (AWPN) to realize on-demand indoor location-based services [1,2]. This paper extends our AWPN to handle huge number of localization requests. In the AWPN, WiFi APs measure received signal strength (RSS) of WiFi signals and send the RSS information to a localization server via a WiFi mesh network. The maximum number of WiFi devices is therefore limited by computational resources on the localization server.咽/epush this limit by introducing a new load balancing technique: we use MapReduce computation framework and perform map operations on APs and reduce operations on localization servers・We also utilize a network router for shu田e operations using network address translation to avoid AP firmware update to adopt changes of the number of localization servers.VIIノe implemented and evaluated our load balancing technique to demonstrate that the load balancing evenly distributes localization calculations to multiple localization servers with approximately 25 % variations.