On the Choice of a Model for Representing Data Flow Parameters in a Digital System

Abstract

The paper considers the problem of analyzing the characteristics of a digital control system as a black box based on the input and output flows characteristics. The Network calculus method is proposed as a system model. The method makes possible to describe the system as a set of deterministic constraints or arrival curve on the flow. The paper proposes to extend the Network Calculus for the case when the flow is described by more than two parameters and extend Network Calculus for multidimensional feature space. An example the feature space may be consist from a packet sequence number, packet size and packet arrival time. The multidimensional flow might be described in a framework of the Network Calculus by some multidimensional arrival cure used to compute boundaries on the flow behavior and characteristics. The proposed extension of the network Calculus makes possible to model a wider class of systems, for example, considering non-uniform data flows.