Repeated-Game Modeling of Multicast Overlays

10 years 7 months ago
Repeated-Game Modeling of Multicast Overlays
— This paper studies multicast application overlay networks in a repeated-game framework. In these overlays, users have both the motivation and the means to alter their position in the overlay tree. We introduce a repeated-game model of user behavior that captures the practical tradeoff between a user’s short-term desire for quality and long-term desire for the network’s continued existence. We simulate overlay treeformation protocols with this model to study their robustness to selfish users. We show that this model can explain user cooperation and provide insight into how overlay systems scale in the absence of heavyweight mechanisms or identity systems. We also use the model to derive practical guidance on how to make multicast overlay protocols more robust to selfish users.
Mike Afergan, Rahul Sami
Added 11 Jun 2010
Updated 11 Jun 2010
Type Conference
Year 2006
Authors Mike Afergan, Rahul Sami
Comments (0)