Its not quite that simple and the other factors are not negligible by any means. The recommended approach would be to download Wireshark, capture the traffic and analyse it in the Throughput Graph.
You can adapt the Wireshark setup from this: http://www.dslreports.com/faq/15888 Regards, Mark McCarron Date: Thu, 2 May 2013 11:17:45 -0700 From: engrsajjadsaf...@yahoo.com Subject: Re: [Discuss-gnuradio] Calculating the delay of TCP link. To: mark.mccar...@live.co.uk Actually i want to calculate the delay using the formula like Assuming all other factors as negligible.Here we have 1500 byte tcp header, which in bits is 1500*8/50KHz. Here the R is in kHz but to use this formula we have to have R in Bits per second. Is my way of calculating is right from this approach or not? Best Regards,SAJJAD SAFDAR From: Mark McCarron <mark.mccar...@live.co.uk> To: Sajjad Safdar <engrsajjadsaf...@yahoo.com> Sent: Wednesday, May 1, 2013 12:51 AM Subject: RE: [Discuss-gnuradio] Calculating the delay of TCP link. The MTU will only tell you if there is fragmentation. In packet switched networks, there can be delays for any number of reasons that are not entirely predictable. For example, assume someone is watching a video, using VOIP, downloading, etc. These can place heavy load on a switch, router or hub and saturate buffers delaying your packets and reducing throughput. Other factors such as QoS or traffic shaping can alter things. Then you have cosmic rays, bad wires, failing circuitry, etc. Then on a PC the network stack itself can be a source of delays as this is implemented in software a dependent on the scheduler and what else is happening in the machine. Trying to monitor all this, only places additional load on these systems a skews your results. The best you can do is attempt to define an average and identify the worst case scenario. Aiming between these two figures will normally provide you with a robust service that exceeds expectation. Regards, Mark McCarron Date: Tue, 30 Apr 2013 12:40:07 -0700 From: engrsajjadsaf...@yahoo.com Subject: Re: [Discuss-gnuradio] Calculating the delay of TCP link. To: mark.mccar...@live.co.uk Hi,Is it any way to calculate using the MTU size of TCP packet and the sampling rate, like a mathematical approach using formulas. Best Regards,SAJJAD SAFDAR From: Mark McCarron <mark.mccar...@live.co.uk> To: Sajjad Safdar <engrsajjadsaf...@yahoo.com> Sent: Tuesday, April 30, 2013 12:33 AM Subject: RE: [Discuss-gnuradio] Calculating the delay of TCP link. Calculating delay is complex. If you just want to know the average time between hosts on an IP network, then use the Ping command. It has a RTT value in ms. Just remember that on a packet switched network, this can vary but is typically under 1ms in a local environment. Similar delays exist throughout the receive chain and processor, which are virtually impossible to measure accurately. Accurate measurements like for radar, or bearings are impossible without some form of time-stamp at the receiver and that would require an atomic clock chip. Regards, Mark McCarron Date: Mon, 29 Apr 2013 12:01:57 -0700 From: engrsajjadsaf...@yahoo.com To: discuss-gnuradio@gnu.org Subject: [Discuss-gnuradio] Calculating the delay of TCP link. Hi,I am sending audio at 50 kHz sample rate via TCP sink from host A to other host B. The host B is connected via router in same network. How can i calculate the time delay from host A to host B via this TCP link. Best Regards,SAJJAD SAFDAR _______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
_______________________________________________ Discuss-gnuradio mailing list Discuss-gnuradio@gnu.org https://lists.gnu.org/mailman/listinfo/discuss-gnuradio
