Equivalent Gibbs Duhem law in network queue

Abstract

High speed networks are highly dynamical systems. Large deviation techniques have been employed to characterize their loss probability. Large deviation is also used to calculate shadow prices of quality of service of a scalable network queue with respect to buffer size B, capacity C and arrival rate N.In this paper we prove that these shadow prices are independent of (B,C,N) parameters. Gibbs Duhem law of thermodynamics is described. Equivalent Gibbs Duhem type law in scalable network queue is derived. Application of equivalent network Gibbs Duhem law is discussed. © IFIP International Federation for Information Processing 2005.