[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Bandwidth Delay Product for High Speed Links and ns2



Hi,

I've got a simple TCL script which comprises 4 nodes (a full TCP source, a
full TCP sink and a bottleneck link).  The data rates are 1.544 Mbps and
the one-way delay is 3 ms, packet size is 1000 bytes.  I wanted to test the
behavior of ns2 with respect to the BDP.

The Bandwidth Delay Product for this link = (1.544 Mbps * .006 sec)/8 =
1158 bytes.

The Question:  It takes the ns2 simulation 6 packets at the TCP source and
intermediate links to "keep the pipe full".  It should only take a maximum
of 2.  Can anyone suggest why this is happening?  Does the TCL script need
to be more intelligent (e.g., specify cwnd, send and receive buffers,
etc.).  I must be missing something.

I can supply the script if desired.

Thanks,
Chuck