Mapping link SNRs of real-world wireless networks onto an indoor testbed

Network simulation packages such as NS-2 and OPNET have been shown to be a limited option for cross-layer experimentation in wireless networking because they cannot faithfully capture the propagation and interference characteristics of wireless channels [1]. Recent research on network cross-layer optimizations further raises this concern due to the close interaction between physical layer feedback and higher layer protocols. To overcome this shortcoming, wireless testbeds have been used wherein novel protocols and application concepts can be assessed in a realistic environment under controlled and repeatable conditions. Since average signal-to-noise-ratio (SNR) often determines the performance of a wireless link, our goal is to seek link SNR mapping methods that replicate real-world link SNRs onto an indoor testbed. Specifically, we devise and assess link SNR mapping methodologies for two different applications: hierarchical networks with a fixed access point (AP), and mesh networks. For the AP-based networks, we employ the minimum weight matching algorithm to minimize the root-mean-square (RMS) mapping error between the testbed and real-world SNRs. For the mesh networks, to avoid the technical difficulties inherent in “forward mapping”, we develop a “reverse mapping” method by which we turn a testbed configuration with specified link SNRs into a real-world configuration. By inducing the link gain difference between the testbed and the real-world distance-dependent path loss to have a log-normal distribution, a very close approximation to real-world shadow fading is achieved. We present results for a variety of indoor and outdoor real-world scenarios to demonstrate the generality of our method.