Abstract—Online games have now become significant contributors to Internet network traffic. As such, research effort is now devoted to the analysis and modeling of this network traffic to aid network designers to provision for gaming traffic in the networks they design. While many research studies have performed measurements on gaming network traffic, the results are game-specific and have not been unified into a framework. In general, network traffic can be represented by packet size and packet interarrival time parameters; therefore, a synthetic model can be built to deal with these two parameters as traffic features. In this paper, we present a technique for constructing a Hierarchical Hidden Markov Model that provides a packet level statistical model for First Person Shooter (FPS) gaming network traffic, which allows generation of traffic for various numbers of users through different game states. The proposed solution has been implemented for Counter-strike and Quake as two of the most popular online FPS games. The results derived from the models have then been used to successfully predict certain game related statistics.
Index Terms—Network traffic generation, gaming traffic simulator, game state modeling.
B. Hariri was with the Distributed and Collaborative Virtual Environment Research (DISCOVER) Laboratory, School of Electrical Engineering and Computer Science, University of Ottawa, Canada. She is now with Google Inc. (e-mail: behnoosh@google.com).
S. Shirmohammadi is with the Distributed and Collaborative Virtual Environment Research (DISCOVER) Laboratory, School of Electrical Engineering and Computer Science, University of Ottawa, Canada (e-mail: shervin@eecs.uottawa.ca).
[PDF]
Cite:Behnoosh Hariri and Shervin Shirmohammadi, "A Statistical Network Traffic Model for First-Person Shooter Games," Journal of Advances in Computer Networks vol. 2, no. 2, pp. 100-105, 2014.
