A matrix interpretation of network calculus

Network calculus theory uses arrival curve and service curve to calculate deterministic bounds of performance parameters. These two notions are defined by min-plus convolution. If we consider the flows in network as vectors, then the arrival and service curve can be defined by matrices multiplication. Thus we can us idempotent matrix analysis which has been a well studied theory to study the theoretical properties of network calculus, thus we provide an goof theoretical background of network calculus theory. Another merit of matrix method is that the min-plus closure of some function is easy to obtain by a matrix method.