[e2e] concurrent TCP connections - how likely are they?

[Duke, Martin]

Mark Allman, "A web servers view of the transport layer'," ACM SIGCOMM
Computer Communication Review, Vol. 30, No. 5 (2000), pp. 10-20.  

Martin

-----Original Message-----
From: Michael Savoric [mailto:savoric at ee.tu-berlin.de] 
Sent: Wednesday, October 22, 2003 8:59 AM
Cc: end2end-interest at postel.org
Subject: [e2e] concurrent TCP connections - how likely are they?


Hi,
I am searching for investigations at huge (WWW) servers to
find
out how likely is it to have more than a single TCP
connection
to the same destination or the same subnetwork at a given
time.

* Do you know some links or papers with regard to this
topic?

* Or is it possible to get (anonymized) traces from such servers
  that I can make my own investigations?


The background is:
I have developed a common congestion controller for TCP connections in
ns-2 and a Linux implementation has been written by one of my students.
Both simulation and measurement results show that such a common
congestion controller reaches (large) performance gains. In addition,
together with some colleagues I have analytically shown that my common
congestion controller is not more aggressive to the network than
standard TCP in its "best case" (Proceedings of ITC 18, September 2003,
Berlin).

What I need now is citeable material to motivate the usage
of a common
congestion controller in the Internet.

Best regards,
Michael Savoric

======================================================
E-Mail:		savoric at ee.tu-berlin.de

Phone:		(+49 30) 314-23840

Fax:		(+49 30) 314-23818

Postal address:	Technical University Berlin
		Telecommunication Networks Group (TKN)
		Einsteinufer 25, 10587 Berlin
======================================================