Sharing the “cost” of multicast trees: an axiomatic analysis

Reviewer: Cecilia G. Manrique

Gone are the days when the Internet was the domain of the military and academics. It is now dominated by business. This commercialization has made issues of cost recovery and profit incentives relevant to discussions about the future of the Internet. The allocation of costs among users, that is, the assignment to various users of some measure of the network resources that they are consuming, will play an important role in future networks. The authors use equations, theorems, and axioms to try to show how one would split the cost of a single data flow among many receivers. Proofs of the theorems are given in an appendix. There are various ways to allocate costs. The authors prefer an axiomatic approach that tries to address the desirable properties of cost allocation schemes, including anonymity, fairness, and equity. They look at the pros and cons of various approaches, and favor what they call the “one-pass” approach. In their analysis, they explain various models and schemes. In the equal tree split scheme, one can divide the total cost equally among all receivers, but this scheme does not distinguish between receivers far from the source and those close to the source. With the standalone axiom, every receiver is guaranteed that joining a group can never cost more than the individual unicast cost. In the “sharing is good” model, no free riders (members who do not pay their fair share) are allowed. This model requires that the allocations decrease as more members join. The authors recognize that their original model addresses only the issue of how to share costs between equivalent users in different locations. The reader is left with the same question that the authors have to grapple with: having multiple quality of service (QoS) levels raises the issue of how to share the cost among several members who are in the same place but request different QoS levels. The paper, though complicated by the mathematical explanations, sheds light on an important aspect of the Internet, which is being debated by various parties interested in the advances in this technology.