Re: Traceroute versus other performance measurement

[Kai Schlichting]

At Wednesday 01:54 PM 11/29/00, smd@clock.org wrote:


>| At the expense of using all their bandwidth ttcp should provide some
>| reasonable measure of performance under those circumstances.
>
>ttcp has the disadvantage of invariant 5-tuple, which is germane to
>this problem.
>
>        Sean.

While this is moving rapidly towards a rather marginal topic, the
above struck me. And Google says for "tuple ttcp" :
http://www.cs.umn.edu/classes/Spring-1999/csci5212-nh/toc.html

Do I understand it correctly that the "invariant 5 tuple" refers to
5 non-variable aspects of this performance test, being namely:

- tcp or udp or icmp(?) protocol
- local IP address
- local protocol port or message type
- remote IP address
- remote port port or message type

E.g.: performance measurements will gravitate towards a result not
representative for the totality of all traffic, as the 5 variables
are rather fixed, or at the least fixed for the duration of the test ?

Has intelligent switching and forwarding (Layer-4 and up) started to
make such a significant impact on end-to-end performance that all
traditional performance tests must necessarily be flawed ?
(there must be a reason why my MRTG graphs of HTTP GET's to websites
 are so much more consistent than icmp pings...)



--
"Just say No" to Spam                                     Kai Schlichting
New York, Palo Alto, You name it             Sophisticated Technical Peon
Kai's SpamShield <tm> is FREE!                  http://www.SpamShield.org
|                                                                       |
LeasedLines-FrameRelay-IPLs-ISDN-PPP-Cisco-Consulting-VoiceFax-Data-Muxes
WorldWideWebAnything-Intranets-NetAdmin-UnixAdmin-Security-ReallyHardMath