Traditional system design assumes that all participants behave according to the intentions of the system architects. In reality, computer networks are heterogeneous, dynamic and distributed environments managed by multiple administrative authorities and shared by users with different and competing interests. Recently, there has been growing interest in using tools from Game Theory (GT) and Mechanism Design (MD) to tackle incentive-related problems in these complex environments.  For these methods to be successful in practical networked systems, it is vital to understand and incorporate realistic models and constraints for such central system properties as player types and strategies, scalability, asynchronicity, observability, verification, and frequency & time scale of interactions.

The goal of this workshop is to promote an exchange of ideas on the true applicability, range and validity of game-theoretic and economic models for analysis and design of Internet and Internet-based systems.  We hope that this will lead to:

1. More awareness within the SIGCOMM community of incentive-based design methodology, and techniques for applying it.
2. A better understanding among game theory researchers of the pragmatic requirements and constraints facing real networks, guiding theoretical research towards more practical system engineering scenarios. 
3. Strengthened research focus in the area, which we hope to achieve by bringing the two communities to agree on a list of specific sub-areas in game theory that are relevant and may be able to capture the characteristics, constraints, and design objectives of networked systems.   
   

This CFP solicits technical or position papers of two types:

Incentives in Practice: Papers that focus on practicality and realistic applications of Game Theory to networked systems. This includes:

• Potential roles for Game Theory and Mechanism Design in networks; e.g. maximizing utilization, countering misbehavior, analyzing performance impact of selfish users, increasing robustness.
• Identifying practical games and incentive issues that arise in Internet-based systems
• Applications of incentive mechanisms (GT, others) to Internet-based systems (P2P, SPAM, Security, Routing, Peering, etc.)
• Methods for creating incentives in distributed systems: reputation, payment, control, legal systems, etc.
• Critiques of prior work and its relevance 

Models & theoretical results for networked systems: Papers that assess the applicability of specific models and/or results in Game theory and Mechanism Design to distributed and networked systems, or outline new approaches to analyzing these problems. This includes:

• Models that capture the distributed nature of the system, asynchronicity, lack of a centralized trusted entity, existence of a large and dynamic number of players, etc. 
• Relevance of repeated games, evolutionary game models, various solution concepts, rationality models, etc. to networked systems
• Relevant possibility and impossibility results.
• Learning and networked agents
• Sensitivity and robustness of GT models
• Distributed Algorithmic Mechanism Design

Submission Instructions

Submissions should be no more than 8 pages in length with 10pt fonts or larger.  Register your paper at the submission page by April 9, 2004. Submit papers in PDF or Postscript format at the submission page by April 19 , 2004.

Important Dates:

   Registration deadline:  April 9, 2004
   Submission deadline: April 19, 2004
   Notification: May 26, 2004
   Camera ready manuscript: June 16, 2004
   Workshop date: September 3, 2004

Workshop Organizers

Dina Katabi, MIT
Rahul Sami, MIT

Organizing and Program Committee

John Chuang, U. C. Berkeley
Jon Crowcroft , Cambridge University
Peyman Faratin, MIT
Dina Katabi, MIT
Peter Key, Microsoft
Kevin Lai, HP Labs
David Parkes, Harvard University
Balaji Prabhakar , Stanford University
Tim Roughgarden, U.C. Berkeley
Rahul Sami, MIT
Scott Shenker, ICSI & U.C. Berkeley
John Wroclawski, MIT