Online gaming has become one of the most profitable businesses on the Internet, with its market share has surpassed those of movie and music industries. With an everlasting interest in computer gaming since childhood, our researching on network gaming dates back from year 2003 with a goal to provide a scalable, efficient, and secure gaming infrastructure for all kinds of network games. To make this line of researches solid and practical, we have been collaborating with a number of game companies, which provide their domain knowledge, real-life operation logs, and funding for research purposes.
Network Game System Design
Traffic Analysis
To gain a better understanding of game traffic, we analyzed a 1,356-million-packet trace from a sizeable MMORPG called ShenZhou Online [1]. This work is, as far as we know, the first formal analysis of MMORPG server packet traces. The results revealed that MMORPG traffic is very different from the traffic of dominant Internet applications, such as web surfing. We identified the following important properties of MMORPG traffic: 1) tiny packets, 2) periodicity, 3) temporal dependence of packet arrivals within connections and aggregate traffic, 4) self-similarity, and 5) heavy-tailed session duration. We have offered physical explanations for the observed traffic characteristics based on the following idiosyncratic user behavior in MMORPGs: 1) the diversity of user behavior, 2) the clustering nature of user inputs, 3) the flash crowd effect, 4) the practice of team play, and 5) heavy-tailed active/idle user activity periods.
"Game Traffic Analysis: An MMORPG Perspective," Kuan-Ta Chen, Polly Huang, and Chin-Laung Lei, Elsevier Computer Networks, Vol. 50, No. 16, pp. 3002--3023, November, 2006. [web | paper]
"Game Traffic Analysis: An MMORPG Perspective," Kuan-Ta Chen, Polly Huang, Chun-Ying Huang, and Chin-Laung Lei, Proceedings of ACM NOSSDAV 2005.
Transport Protocols
Although online games are becoming increasingly popular as well as a highly profitable Internet business, there is still a fundamental design question: Which is the best transport protocol for online gaming? TCP, UDP, or any other protocol? We are the first to evaluate whether TCP is suitable for online games [1], and have proposed novel transport strategies [2] for this genre of applications. Our analysis of a trace collected from a TCP-based MMORPG called ShenZhou Online indicates that TCP is unwieldy and inappropriate for MMORPGs. To provide better gaming experience, our proposed transport strategies assign game packets with appropriate levels of transmission guarantee depending on the requirements of the packets' contents. Via real-life game traces from Angel's Love, we proved that our strategies effectively raise the satisfaction levels of the game players.
"An Empirical Evaluation of TCP Performance in Online Games," Kuan-Ta Chen, Chun-Ying Huang, Polly Huang, and Chin-Laung Lei, Proceedings of ACM SIGCHI ACE 2006. [web | paper]
"On the Challenges and Designs of Transport Protocols for MMORPGs," Chen-Chi Wu, Kuan-Ta Chen, Chih-Ming Chen, Polly Huang, and Chin-Laung Lei, Multimedia Tools and Applications (special issue on Massively Multiuser Online Gaming Systems and Applications), Vol. 45, No. 1, pp. 7--32, 2009.
Implications from User Interaction
While psychologists analyze network game-playing behavior in terms of players' social interaction and experience, understanding user behavior is equally important for network researchers, because how users act determines how well network systems, such as online games, perform. From the ShenZhou Online trace [1,2], we find that the dispersion of players in a virtual world is heavy-tailed, which implies that static and fixed-size partitioning of game worlds is inadequate. Neighbors and teammates tend to be closer to each other in network topology, which implies that message delivery between the hosts of interacting players can be faster than between that of unrelated players. We also find that participants who have a higher degree of social interaction tend to play much longer, and players who are closer in network topology tend to team up for longer periods. This suggests that game designers could increase players' loyalty by encouraging or even forcing team playing.
"Network Game Design: Hints and Implications of Player Interaction," Kuan-Ta Chen and Chin-Laung Lei, Proceedings of ACM NetGames 2006. [web | paper]
"Design Implications of Social Interaction for Online Games," Kuan-Ta Chen and Chin-Laung Lei, Proceedings of IFIP ICEC 2006 (poster).
Cloud Technology
Nowadays MMORPG servers tend to be overly provisioned because 1) such games do not have standard architectures thus dedicated hardware is assumed; 2) MMORPGs normally adopt a "sharded design" to resolve the scalability challenges of content production and workload distribution; and 3) a game is commonly deployed in geographically distributed data centers to protect gamers from excessive network latencies. Therefore, an operator needs to deploy dedicated hardware for each game in each data center, even though hardware utilization is low. To address these issues, we have firstly proposed a zone-based server consolidation strategy for MMORPGs [1], which exploits the unique spatial locality property of players' interactions, to cut down the games' considerable hardware requirement and energy use. We evaluate the effectiveness of our strategy based on a nine-month trace from a popular MMORPG World of Warcraft. The evaluation results showed that, with a per-hour dynamic zone reallocation policy, the server number required can be reduced by 52% and the total energy consumption can be reduced by 62%, while the user-experienced latency remains undegraded.
"Is Server Consolidation Beneficial to MMORPG? A Case Study of World of Warcraft," Yeng-Ting Lee and Kuan-Ta Chen, Proceedings of IEEE CLOUD 2010. [web | paper]
"Measuring The Latency of Cloud Gaming Systems," Kuan-Ta Chen, Yu-Chun Chang, Po-Han Tseng, Czhun-Ying Huang, and Chin-Laung Lei, Proceedings of ACM Multimedia 2011. [web | paper]
Network Game Cheating Issues
Game Bot Detection
Despite online games' success, one of the games' greatest challenges is the increasing use of game bots, i.e., auto-playing game clients. The use of game bots is considered unsportsmanlike and is therefore forbidden. To keep games in order, game police, played by actual human players, often patrol game zones and question suspicious players. This practice, however, is labor intensive and ineffective. As a pioneer to this problem, we have proposed a few alternatives to automatically detect the use of game bots without intervening legitimate users' gaming experience. Our approaches are based on network traffic generated by game clients [1], the movement trajectories of the player-controlled avatars [2,3,4], the trading behavior between avatars [5], and how players respond to harmonic stimuli [6], respectively, which are appropriate in different scenarios for the best performance.
"Identifying MMORPG Bots: A Traffic Analysis Approach," Kuan-Ta Chen, Jhih-Wei Jiang, Polly Huang, Hao-Hua Chu, Hao-Hua Chu, Chin-Laung Lei, and Wen-Chin Chen, EURASIP Journal on Advances in Signal Processing, Article ID 79715, 2009. [web | paper]
"Game Bot Detection via Avatar Trajectory Analysis," Hsing-Kuo Pao, Kuan-Ta Chen, and Hong-Chung Chang, in IEEE Transactions on Computational Intelligence and AI in Games, vol 2, no 3, Sep, 2010.
"Game Bot Detection Based on Avatar Trajetory," Kuan-Ta Chen, Andrew Liao, Hsing-Kuo Kenneth Pao, Hao-Hua Chu, Proceedings of IFIP ICEC 2008. [web | paper]
"Game Bot Identification based on Manifold Learning," Kuan-Ta Chen, Hsing-Kuo Kenneth Pao, and Hong-Chung Chang, Proceedings of ACM NetGames 2008. [web | paper]
"Detection of MMORPG Bots Based on Behavior Analysis," Ruck Thawonmas, Yoshitaka Kashifuji, and Kuan-Ta Chen, Proceedings of ACM ACE 2008.
"Bot Detection in Rhythm Games: A Physiological Approach," Ruei-Min Lin, Hwai-Chung Ho, and Kuan-Ta Chen, Proceedings of ACM SIGCHI ACE 2011. [web | paper]
Account Hijacking
Account hijacking is also considered one of the most serious security problems in online games. A hijacker normally takes away valuable virtual items from the stolen accounts, and trades those items for real money. Even though account hijacking is not uncommon, there are currently no general solutions to determine whether an account has been hijacked. The game company is not aware of a hijack unless it is reported by the victim. However, it is always too late---the victim's valuables have already been stripped off. To address this issue, we have proposed a new biometric for human identification based on users' idle time distributions [1]. The evaluation shows that our scheme achieves higher than 90% accuracy with a 20-minute detection time given a 200-minute history size.
"User Identification based on Game-Play Activity Patterns," Kuan-Ta Chen, Li-Wen Hong, Proceedings of ACM NetGames 2007. [web | paper]
Network Game User Behavior Analysis, Modeling, and Prediction
Temporal Behavior Analysis
Most revenues of the game industry come from the sale of subscriptions and virtual items, especially to loyal "hardcore" players who would stay in a game for more than a year. Understanding the players' behavior and how long they might stay in the game is hence vital to game operators. If a player's departure is predictable, measures can be taken to prevent that from happening. We have developed a practical scheme for predicting player's unsubscription [1,2,3], which allows us to predict two months prior with a compound accuracy of over 80%. We have shown that our scheme is generalizable across different games and can be applied to avatar usage predictions.
"World of Warcraft Avatar History Dataset," Yeng-Ting Lee, Kuan-Ta Chen, Yun-Maw Cheng, and Chin-Laung Lei, Proceedings of ACM Multimedia Systems 2011. [web | paper]
"An Analysis of WoW Players' Game Hours," Pin-Yun Tarng, Kuan-Ta Chen, and Polly Huang, Proceedings of ACM NetGames 2008.
Spatial Behavior Analysis
Understanding of player's behaviors can help game designers improve the design of game play mechanisms and enable personalized game content. We have conducted several studies on game avatars' movement patterns, including 1) automatic landmark identification in the virtual worlds where landmarks are those locations game players frequently visit [1,2]; 2) clustering of game players based on their landmark visiting patterns [1,3]; and 3) summarization and clustering of game players' in-game area revisitation patterns [4,5,34]. These works provide useful tools for game designers and operators to visually examine the macro-level movement behavior of their players for improvement and possible automatic adaption of game content.
"Detection of Landmarks for Clustering of Online-Game Players," Ruck Thawonmas, Masayoshi Kurashige, and Kuan-Ta Chen, International Journal of Virtual Reality, Vol. 6, No. 3, pp. 11--16, September 2007.
"Detection of Landmarks for Clustering of Online-Game Players," Ruck Thawonmas, Masayoshi Kurashige, and Kuan-Ta Chen, Proceedings of IFIP ICEC 2007.
"Comparison of User Trajectories Based on Coordinate Data and State Transitions," Junichi Oda, Ruck Thawonmas, and Kuan-Ta Chen, Proceedings of IEEE IIHMSP 2009.
"Analysis of Area Revisitation Patterns in World of Warcarft," Ruck Thawonmas, Keisuke Yoshida, Jing-Kai Lou, and Kuan-Ta Chen, Proceedings of IFIP ICEC 2009.
"Analysis of User Trajectories Based on Data Distribution and State Transition: A Case Study with a Massively Multiplayer Online Game -- Angel Love Online," Ruck Thawonmas, Junichi Oda, and Kuan-Ta Chen, Proceedings on GAMEON 2009.
"Comparison of User Trajectories Based on Coordinate Data and State Transitions," Junichi Oda, Ruck Thawonmas, and Kuan-Ta Chen, Proceedings of IEEE IIHMSP 2009.
Social Network Analysis
We are also interested in how players interact with each other in games. We analyzed a real-life online game's social network and identified that the players' in-game social behavior highly correlate with their game-play activities and that the former is a predictor for a player's future activities. In [1], we proposed a method for the efficient computation of the vertex-betweenness centrality index of a node in a social network. While the traditional way of computing the index is costly, i.e., in the order of O(VE), our estimate method requires only O(b2V) time, where b is the average node degree of the network, and retains the accuracy of the index.
"Estimating Vertex-Betweenness Values Based on Local Dynamic Information for Online Game Social Networks," Jing-Kai Lou, Shou-de Lin, Kuan-Ta Chen, and Chin-Laung Lei, ASONAM 2010.