Measuring 'time interval of transfer' ofpackets
carter at qosient.com
carter at qosient.com
Fri Jul 21 05:58:46 EDT 2006
Hey VAIBHAV,
You do what you're after by having argus in both machines. If the timestamps are synchronized, then you just subtract them and you get one-way delay.
If the timestamps are not synchronized, you cannot calculate one-way delay, but you can calculate a more accurate round trip time by subtracting the "dur" values from the two argus outputs.
Carter
------Original Message------
From: vaibhav_kaware at persistent.co.in
Sender: argus-info-bounces at lists.andrew.cmu.edu
To: Argus
Sent: Jul 17, 2006 2:36 AM
Subject: [ARGUS] Measuring 'time interval of transfer' ofpackets
Hi,
I intend to measure the time of transfer of a given packet
from source to the destination.
But each packet captured/recorded has a single time stamp,
that of arrival on the client machine. This only marks a
single point on the timeline. Whereas, duration implies an
interval bounded by two points.
Is it possible to get the duration of travel of the packets
using Argus? If yes, how?
Alternatively, is it possible to get the arrival timestamps of
the packets, 'of the other end'?
If not, what could I employ to perform the measurements?
-VAIBHAV
Please read further if further clarification required.
-----------------------------------------------------------
e.g.
If there is a webpage small enough that gets transferred from
the server to the client machine within one packet, then there
is no way of knowing how long it took for the transfer from
server to the client, looking just at the arrival time-stamp
on the packet.
e.g.
If a 'file' transfer (download) takes place in 'n > 1' number
of packets, then seldom will all the packets arrive in an
uninterrupted queue at the client's machine. In that case, it
wont be possible to conclude about the download time of the
'file' as a whole.
e.g.
When I send a message, say, using the instant messanger,
there is no way for me to know how much time it took for my
message to reach the other user, as all the timestamps on the
packets are the ones that belong to 'my' machine. What about
the arrival
Carter Bullard
QoSient LLC
150 E. 57th Street Suite 12D
New York, New York 10022
+1 212 588-9133 Phone
+1 212 588-9134 Fax
More information about the argus
mailing list