Title :
The revolution of StarCraft network traffic
Author_Institution :
Dept. of Math. & Comput. Sci., Gustavus Adolphus Coll., St. Peter, MN, USA
Abstract :
StarCraft II uses a client-server architecture while StarCraft I uses a peer-to-peer architecture. StarCraft I clients utilize more outgoing bandwidth for each additional player in the game, while StarCraft II clients maintain constant client-to-server bandwidth regardless of the number of players. StarCraft I and StarCraft II exhibit distinct packet size and inter-packet size distributions. For StarCraft II, we model inter-packet time distributions for both client-to-server and server-to-client traffic regardless of the number of players. The packet size distribution for client-to-server traffic also remains constant regardless of the number of players while the packet size distribution for server-to-client traffic shifts with regard to the number of players. We segmented each of these distributions, and fit each segment with a uniform, beta, gamma or triangular distribution, constructing the model for StarCraft II network traffic.
Keywords :
client-server systems; computer games; delays; digital simulation; jitter; peer-to-peer computing; quality of experience; sport; telecommunication traffic; StarCraft I game players; StarCraft II game players; StarCraft network traffic characteristics; casters; delay condition; e-sports games; electronic sports; emulation tool; experience degradation; jitter condition; loss condition; network conditions; online games; professional matches; professional players; quality-of-experience; real-time strategy game; routine matches; simulation tool; spectators; Emulation; Games; Mathematical model; Peer to peer computing; Servers; Standards; Throughput;
Conference_Titel :
Network and Systems Support for Games (NetGames), 2012 11th Annual Workshop on
Conference_Location :
Venice
Print_ISBN :
978-1-4673-4576-7
Electronic_ISBN :
2156-8138
DOI :
10.1109/NetGames.2012.6404022