Automatic retrieval of network traffic data for analysis of network-in-world action relations in MMOGs

Abstract

Massive multiplayer online games (MMOGs) are increasingly popular because they provide entertainment, numerous opportunities for socialization, and the ability for users to make money. Cornerstone to MMOGs is the underlying network traffic between MMOG clients and servers; understanding this traffic is important for application developers trying to optimize game performance and for ISPs trying to provide a better quality of service for their customers. This paper goes a step forward in helping to understand MMOG network traffic as it describes an automated approach to collect traffic samples and map them to specific in-world actions in the MMOG game Second LifeTM. We show the validity of our approach by collecting 10 samples from 8 different actions, characterizing these samples, and grouping samples by action using a k-means clustering algorithm.