From: Jon Webb (jon_webb@binary-one.com)
Date: 09/28/04
Message-ID: <415983AF.70007@binary-one.com> Date: Tue, 28 Sep 2004 08:30:55 -0700 From: Jon Webb <jon_webb@binary-one.com> Subject: Re: Using tcptrace with ns traces
This is a question that should be addressed to the authors of the NS
trace generator. I have done a lot of work (read: a thesis) on
generating real network traces from simulators, in my case Qualnet and
Glomosim, and I can tell you that it is not a simple matter. It is quite
possible that the NS authors didn't get it quite right, or as you point
out, certain kinds of packets that would otherwise be ignored in the
simulator are seen as dups. In fact, in all of these simulators, there
are many events which can look like dups to tcptrace, especially if the
protocols in question are maintaining a packet cache which would most
definitely look like a tcp dup, if not a hardware dup.
Bharath Bhushan wrote:
>Hi all,
>
>1) When the two endpoints are 0.0 and 3.0, the tcp connection between
>them is being recongnized as two separate half-duplex connections. If
>I change the endpoint addresses to (0.0, 3.1) or (0.1, 3.0), it is
>recognized as one complete connection. What could the problem be?
>
>2) The final statistics shows lots of duplicates. When I looked at the
>pread_ns_fulltcp() function, it doesn't seem t care about the
>different kinds of events happening on a packet. So the
>enque/deque/recv etc are all recognized as duplicate packets. Is there
>a way to overcome this?
>
>Thanks in advance
>-- Bharath
>----------------------------------------------------------------------------
>To unsubscribe, send a message with body containing "unsubscribe tcptrace" to
>majordomo@tcptrace.org.
>
>
>
----------------------------------------------------------------------------
To unsubscribe, send a message with body containing "unsubscribe tcptrace" to
majordomo@tcptrace.org.
This archive was generated by hypermail 2.1.7 : 09/28/04 EDT