Iterative computations of bit rate summary statistics at inlet and outlet of a discrete time fluid queue

• Network performance management is facing the challenge of provisioning advanced services with stringent delay and throughput requirements. For this reason, shortage of network capacity implying delay or loss, so-called bottlenecks, have to be identified and to be classified. The latter tasks imply the need for tractable analytical performance models. As we are particularly interested in the traffic flow properties, we focus on a discrete-time fluid flow traffic model. The fluid model considers averages of bit rates on small time scales. These averages can easily be measured, and thus, bit rate statistics for data streams entering and leaving a bottleneck can be obtained. The fluid flow model helps us to obtain analytical estimates on the impact of network entities on data streams and their statistics, i.e., efficiently predict the effects of bottlenecks on data streams. In this work we derive formulae for an iterative calculation of summary statistics such as the bit rate distribution and autocorrelation function. In addition, a corresponding simulation has been implemented, and the simulation results agree closely with the exact solutions. With the aid of the derived formulae we are able to match the model to observations.

• Pappret addresserar ett tidsdiskret flödes-kömodell med fokus på bithastighetsstatistik såsom bithastighetshistogramm och autokorrelation. Flödesmodell, kömodell, tidsdiskret modell, prestandaberäkning

Ämnesord

TEKNIK OCH TEKNOLOGIER  -- Elektroteknik och elektronik -- Telekommunikation (hsv//swe)

ENGINEERING AND TECHNOLOGY  -- Electrical Engineering, Electronic Engineering, Information Engineering -- Telecommunications (hsv//eng)

Nyckelord

Fluid flow model

discrete-time queuing model

bit rate statistics

performance evaluation

Publikations- och innehållstyp

ref (ämneskategori)

kon (ämneskategori)