A fluid-flow model of RIO routers loaded by Markov modulated fluid processes

In the last few years some mechanisms have been proposed to support service differentiation in the Internet, to meet the requirements of different types of traffic flows. More in depth, a congestion control mechanism, named RIO, has been proposed to offer different drop precedence to packets having different priorities. The target of this paper is to provide an analytical fluid-flow model of RIO routers loaded by any kind of traffic sources that can be represented as Markov modulated fluid processes (MMFP) (e.g. video, audio and multimedia sources), and to obtain from the model an analytical evaluation of their steady-state performance.