Toward microeconomic allocation of resources in multi-service overlay networks

Abstract

The main challenge in overlay multicasting is designing self-organizing mechanisms that can be able to exploit the inherent selfishness of the end-user nodes in such a way that the aggregate outcome of the activity of individual nodes behaving toward their own self-interests still leads to maximization of the network's aggregate utility. We believe that the microeconomic theory is a good candidate to investigate this problem. Since each consumer in the economy acts as a selfish utility maximizer, the behavior of each end-host in the overlay network can be mapped to that of a consumer in the economy. We present a competitive economical framework in which a number of independent services are provided to the users by a number of origin servers. Each offered service can be thought of as a good and the origin servers and the users who relay the service to their downstream nodes can thus be thought of as firms of the economy. Also, the end-hosts can be viewed as consumers in the economy. On joining to the overlay network, each end-host is provided with an income and tries to obtain the services. The mechanism tries to regulate the price of each service in such a way that general equilibrium holds. For this property to hold in all generality, it tries to find a vector of prices such that demand of each service becomes equal to its supply. So, all allocations will be Pareto optimal in the sense that maximize welfare of the users.