[e2e] simulation using self-similar traffic

Mark Crovella crovella at cs.bu.edu
Wed Apr 4 09:48:18 PDT 2001


Bill - No screw-up there, that's indeed what typically happens.

Mark

-----Original Message-----
From: end2end-interest-admin at postel.org
[mailto:end2end-interest-admin at postel.org]On Behalf Of Bill Courtney
Sent: Wednesday, April 04, 2001 12:28 PM
To: huang at tik.ee.ethz.ch; shlee at mmlab.snu.ac.kr
Cc: end2end-interest at postel.org
Subject: RE: [e2e] simulation using self-similar traffic


Soo-hyeong & Polly,

If you take the conversation off-line, please share with the list any
conclusions
that you come to. I, for one, would be very interested.

My limited experience with heavy-tailed distributions has always been less
than
satisfactory. I think that because the probability that very long file sizes
will
be drawn is so low (though non-negligible), the sample average file size is
almost
always less than the true average of the distribution. The result I get is
that
the simulated offered traffic is less than the expected offered traffic,
and,
ironically, performance ends up looking better than expected rather than
worse.

Or it could be that I screwed something up. :-)

Regards,
Bill

[snip]
>
> Nevertheless I agree the average link throughput is bounded by the
> bandwidth.  And, the discussion is going a bit too mathematical.
> Perhaps we should take it offline.
>
> Cheers,
> -Polly
>




More information about the end2end-interest mailing list