How to measure the throughput of a network with ipmt Unix command

Ipmt, for IP measurement tools, is a set of tools able to carry out performance measurements on IP networks at the transport layer. tcpmt and udpmt are network performance measurement tools at the transport layer (respectively TCP and UDP). Every second, they generate a throughput report.
By default, they transmit on the discard port of the destination station. This means that, for UDP, the measured throughput at the sender cannot be the effective transmitted throughput since UDP is not reliable.

How to use them :
# tcpmt [-p port] [-s pktsz] [-n pktnb] host
# udpmt [-p port] [-s pktsz] [-n pktnb] host

Where :
port is the destination port number (discard, by default).
pktsz is the size of sent packets (1024 bytes by default) This parameter is only significant in the UDP context.
pktnb is the packet number to send (by default, press Control-C to stop the transmission).
host is the destination address.
A report is generated every second

• transmission time (in ms since 0h),
• the number of sent packets during the time interval,
• the total number of sent packets,
• the measured throughput during the time interval in kbit/s,
• the average throughput on the last ten seconds,
• the average throughput since the beginning.

tcptarget and udptarget are two programs which measure the trhoughput at the destination. Thay can be launched with no argument and they display the port number on which they are listening. We must use the -p option of tcpmt/udpmt in order to specify a port. As soon as tcptarget/udptarget start to receive packets, they display a report every second. This report is easier to understand than the sender report but it is less detailed. It displays the reception time in seconds with a precision around the millisecond and the throughput in kbit/s.

If you want to measure the behavior of concurrent flows sent on the same target station, Tcp/udptarget are the perfect tools since the arrival times are computed on the same station. Thus you can easily compare the packet arrival times of the different flows without having to synchronize the sending stations.