Mapping MMOFPS over heterogeneous distributed systems

This paper evaluates the effects on performance of taking the heterogeneity of nodes into account, in terms of number of cores, when MMOFPS (Massively Multiplayer Online First Person Shooter) game services are distributed. Two mapping strategies, Non_Heterogeneity_aware and Heterogeneity_aware, are integrated in a system, called OnDeGaS (On Demand Game Service), which is a hybrid between Client-Server and P2P topologies. Through simulation, we show that the Heterogeneity_aware mechanism has more impact in communication costs, but it is more efficient in exploiting the nodes of the P2P area, as it maps players faster and it creates less computing zones with latency values under the acceptable threshold for MMOFPS games.