P2P matchmaking solution for online games

Abstract

Matchmaking players is an important problem in online multiplayer games. Existing solutions employ client-server architecture, which induces several problems. Those range from additional costs associated with infrastructure maintenance to inability to play the game once servers become unavailabe (due to being under Denial of Service attack or being shut down after earning enough profit). This paper aims to provide a solution for the problem of matchmaking players on the scale of the Internet, without using a central server. In order to achieve this goal, the SelfAid platform for building custom P2P matchmaking strategies is presented. After the developer creates a service algorithm defining the matchmaking behavior specific to his/hers case, the SelfAid platform designates a number of player machines to execute the service. Furthermore, the number of designated machines adapts to the demand. SelfAid uses only spare resources of player machines, following the trend of sharing economy. A distributed algorithm is presented and its correctness is proven.