Thank you Marco, i think the problem was tha t i didn't multiply with the previous number of pkts.
Regards. On 7/17/06, Marco Fiore <[EMAIL PROTECTED]> wrote: > > Nicholas, > > I think the right way would be: > > mean = [mean * (total_pkts- > 1) + curr_pkt_delay] / total_pkts > > where total_pkts represents the > number of packets received, including > the current one (i.e., that > determining the curr_pkt_delay contribution). > > Regards, > > Marco > > >---- > Messaggio originale---- > >Da: [EMAIL PROTECTED] > >Data: 17- > lug-2006 3.06 AM > >A: <ns-users@ISI.EDU> > >Ogg: [ns] Calculating Mean end- > to-end delay. > > > > > >Hello, > > > >i have created an agent which allows me to > record the end-to-end delays of > >packets arriving at a specific node. > What i want to do is to calculate the > >mean end-to-end delay at > specific time intervals and later on output it on a > >trace file. > > > >To > calculate the mean end-to-end delay i use the following: > > > >mean= > (mean+curr_pkt_delay)/total_pkts. > > > >I do this every time a new packet > arrives. > > > >Is this the correct way of calculating mean end-to-end > delay or i am totally > >wrong here? > > > >Regards, > > > >-- > >Nicholas > Loulloudes > > > >Postgraduate at Communication Networks and Software. > > > >BSc in Computer Science. > > > > > -- Nicholas Loulloudes Postgraduate at Communication Networks and Software. BSc in Computer Science.
