2nd International ICST International Workshop on Network Simulation Tools

Research Article

FlowMonitor - a network monitoring framework for the Network Simulator 3 (NS-3)

  • @INPROCEEDINGS{10.4108/ICST.VALUETOOLS2009.7493,
        author={Gustavo Carneiro and Pedro Fortuna and Manuel Ricardo},
        title={FlowMonitor - a network monitoring framework for the Network Simulator 3 (NS-3)},
        proceedings={2nd International ICST International Workshop on Network Simulation Tools},
        publisher={ACM},
        proceedings_a={NSTOOLS},
        year={2010},
        month={5},
        keywords={},
        doi={10.4108/ICST.VALUETOOLS2009.7493}
    }
    
  • Gustavo Carneiro
    Pedro Fortuna
    Manuel Ricardo
    Year: 2010
    FlowMonitor - a network monitoring framework for the Network Simulator 3 (NS-3)
    NSTOOLS
    ICST
    DOI: 10.4108/ICST.VALUETOOLS2009.7493
Gustavo Carneiro1,*, Pedro Fortuna1,*, Manuel Ricardo1,*
  • 1: INESC Porto, Faculdade de Engenharia, Universidade do Porto
*Contact email: gjc@inescporto.pt, pedro.fortuna@inescporto.pt, mricardo@inescporto.pt

Abstract

When networking researchers meet the task of doing simulations, there is always a need to evaluate the value of such models by measuring a set of well known network performance metrics. However, simulators in general and NS-3 in particular, require signifcant programming effort from the researcher in order to collect those metrics. This paper reports a contribution for NS-3 consisting of a new ow monitoring module that makes it easier to collect and save to persistent storage a common set of network performance metrics. The module automatically detects all ows passing through the network and stores in a fle most of the metrics that a researcher might need to analyze about the ow, such as bitrates, duration, delays, packet sizes, and packet loss ratio. The value of this module is demonstrated using an easy to follow example. It is also validated by comparing the measurements of a simple scenario with the expected values. Finally, the performance of ow monitoring is characterized and shown to introduce small overheads.