Abstract
|
Enterprise networks provide communication services to business critical applications and must therefore deliver high performance. This demands for thorough monitoring and management of such networks. One-Way Delay (OWD) is an important measurand in this domain, as OWD directly impacts service response time and is an indicator for network congestion. This thesis proposes a novel method for flow-based OWD measure-ment. As it takes easily obtainable flow data as input, there is almost no additional effort and OWD samples are directly obtained from production traffic. The flow-based method is built upon models for flow capturing effects and measurement errors. An online processing method extracts OWD samples from flow data and performs error compensation and quantification based on the models. The device-specific model parameters are given in exporter profiles generated by an offline processing method. The online and offline processing methods have been implemented in a prototype and validated in a large enterprise network scenario.
|
Reference entry
|
Kögel, J.
One-Way Delay Measurement based on Flow Data in Large Enterprise Networks - Communication Networks and Computer Engineering Report No. 111
Dissertation, Universität Stuttgart, Informatik, Elektrotechnik und Informationstechnik, 2013
|