Internet applications with more and more different (real-time) services
are introduced. The reason that real-time services are nowadays
available is due to the fact that higher bandwidth networks offer low
latency. These new (real-time) services are more sensitive to latency
and therefore require lower latency figures. Due to the fact that these
services send real-time traffic it stimulated the network engineers to
focus on the obstacles in developing and providing these services in a
more efficient and scalable way. Since all data-traffic in networks has
latency, network operators want to know how large this latency is.
This project addresses measuring latency in the Versatel Backbone
network and within this network there are a number of specific services
of which the latency was measured during this project. In order to
present a good overview of all the latency measurements between the
POP's in the network is to put the measurement figures in a latency
matrix. The main goal of this project is to develop a latency
measurement architecture for the Versatel Backbone network. A part of
this architecture is a Latency Matrix Application, during this project a
prototype of this application has been produced.
For performing accurate, representative measurement figures the research
of this project focussed on the accuracy, time resolution and
distribution of the measurements. As alternative for the currently used
uniform (periodic) distributed probing also Poisson process distributed
probing for performing the measurements has been studied. Further in
this project, it has been checked if the measurements have influence on
the network performance.