Network Calculus (NC) is a powerful mathematical theory for the performance evaluation of communication systems, since it allows to obtain worst-case performance measures. In communications system modeling, the NC theory is often used to determine Quality of Service (QoS) guarantees, for example of packet-switched systems. In networking systems, the aggregation of data flows plays an important role while modeling the multiplexing scheme. When the multiplexing order is not First in, First out (FIFO), the strictness of the service curve plays an important role. This article deals with problems that arise from the strictness requirement considering aggregate scheduling. The literature reports that the strictness of an aggregated service curve is a fundamental precondition to obtain the individual service curve for a single left-over flow when a node processes multiple input flows in a non-FIFO manner. In many publications, this important strictness property is assumed to be a feature of the service curve only. We will show that, in general, this assumption is not true. In most cases, only the concrete input flow in connection with the service curve allows to decide whether the service curve is strict or non-strict. However, the abstraction from a concrete input flow with an arrival curve as upper bound is not enough to determine the service curve’s strictness. Therefore, to bypass the strict-non-strict problems, we devise theorems to gain guaranteed performance values for a left-over flow. Keywords– Worst-case Communication System Modeling; Network Calculus; Aggregate Scheduling; Strict-non-strict Service Curves; Backlogged Period
[1]
Jens B. Schmitt,et al.
Delay Bounds under Arbitrary Multiplexing
,
2007
.
[2]
Rene L. Cruz,et al.
A calculus for network delay, Part I: Network elements in isolation
,
1991,
IEEE Trans. Inf. Theory.
[3]
Jean-Yves Le Boudec,et al.
Delay Bounds in a Network with Aggregate Scheduling
,
2000,
QofIS.
[4]
Reinhard German,et al.
A Framework for Establishing Performance Guarantees in Industrial Automation Networks
,
2014,
MMB/DFT.
[5]
Volker Sander,et al.
A parameter based admission control for differentiated services networks
,
2004,
Comput. Networks.
[6]
Bruno Gaujal,et al.
Optimal routing for end-to-end guarantees: the price of multiplexing
,
2007,
Valuetools 2007.
[7]
Catherine Rosenberg,et al.
Buffer overflow asymptotics for multiplexed regulated traffic
,
2008,
Perform. Evaluation.
[8]
Florin Ciucu,et al.
Delay Bounds in Communication Networks With Heavy-Tailed and Self-Similar Traffic
,
2009,
IEEE Transactions on Information Theory.
[9]
Ivan Martinovic,et al.
Improving Performance Bounds in Feed-Forward Networks by Paying Multiplexing Only Once
,
2008,
MMB.
[10]
Reinhard German,et al.
Stochastic and deterministic performance evaluation of automotive CAN communication
,
2009,
Comput. Networks.
[11]
G. Rizzo.
Stability and bounds in aggregate scheduling networks
,
2008
.