With the popularity of the Internet, it is logical that computer games have evolved into Internet games. Unlike face-to-face games in which a player can see his or her opponents, Internet game players communicate by transmitting messages over the Web. As a result, the messages may suffer from delay and loss. It is indistinguishable to human players whether a delay is the result of a player thinking or of a congested network. Therefore, cheating is possible by deliberately delaying a decision until one sees the actions taken by other players and pretending the delay was caused by the network.

This paper studies the cheating problem in both centralized and peer-to-peer games. In the centralized version, a server maintains the states of all players and makes state adjustments. In the peer-to-peer model, each player will have a necessary view, if not a global view, of other players’ states for making the moves. A simple protocol for peer-to-peer game play, called Lockstep, is designed to eliminate cheating. Then, an extension, the asynchronous synchronization protocol, is proposed to improve the performance. Finally, the scalability problem is addressed by considering players’ locations.

The kind of cheating studied in this paper will not be a problem when the Internet infrastructure is revamped. Nonetheless, this paper is interesting and clearly written.