Abstract

This paper proposes a framework for analysing admission controls as pricing schemes for shared services. Token bucket admission-control mechanisms are considered as pricing schemes. To analyse the buyer's problem of choosing optimal parameters for token bucket schemes, an important performance metric of token bucket mechanisms is considered, the long-run probability of being denied service, which is equivalent to the threshold-crossing probability of two-sided or one-sided regulated, random walks. For the buyer's problem, the paper gives approximations for these hard-to-calculate metrics, shows that the problem is convex under mild assumptions, and provides a closed-form solution to an approximation of the problem when demand is normally distributed.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call