A Statistical Approach to Cheating Countermeasure in P2P MOGs

Detecting cheaters in peer-to-peer multiplayer online games (MOGs) may be more effective than avoiding the cheats. This is particularly true for fast paced, real-time MOGs, since the classic approaches to prevent time cheats usually complicate the communication protocols among peers. This results in a loss of responsiveness among players during the game session. We present a cheating detection scheme, named AC/DC (Algorithm for Cheating Detection by Cheating), that faces a specific time cheat, i.e., the look-ahead cheat. The idea is that of exploiting a counterattack approach to determine if the suspected node waits for messages coming from other peers before generating its cheated events. During the counterattack, the controlling node delays the transmission of newly generated events to the suspected node. This allows to determine if the suspected node waits for these events before generating its own ones. In this paper, a novel version of the AC/DC approach is presented which exploits a control scheme that measures the statistical correlation between the additional delay employed by the controlling peer and the network latencies from the controlled peer. Simulation results confirm the viability of the approach.