Collaborative Information Revelation and Collusion in "Word-of-Mouth" Access Selection Schemes

"Word-of-mouth", a novel scheme for sharing the service quality experienced with different local networks, is presented and evaluated as a distributed solution for performing user-initiated access selection in a multi-system scenario. If only partial information, about the different access alternatives, is available, by sharing and processing individually discovered information, terminal agents can collectively reveal otherwise concealed information and perform more informed access selection decisions. Introducing collusion mechanisms, between low performing networks and terminal agents, we investigate the robustness, of our distributed mechanism, to the dissemination of false information. In a scenario in which user-deployed APs have different degrees of backbone capacity limitation we investigate, if local operators can attract more traffic, by monetarily bribing terminal agents to transfer false information. Furthermore, we evaluate, under which conditions the prediction capability of our scheme can still support roaming access selection decisions that improve user service appreciation. The results show that if "enough" information, can be collected by independent sources, "economically rational" low performing APs can be effectively prevented from both concealing information and/or attempting to influence their reputation through collusive behavior.