Re: [colorforth] End-to-End Internet Packet Dynamics

["Samuel A. Falvo II" <kc5tja@xxxxxxxx>]

On Tuesday 27 April 2004 10:22 pm, Mark Slicker wrote:
> I question why they didn't use a data rate for flow control. Also all

Because data rate isn't flow control, in the sense of achieving a 
reliable link.  At best, it's flow throttling and bandwidth control.  
Sending data slower does not guarantee a reliable delivery of a packet.

> the ACKs returning to the server seems wasteful, especially in

Wasteful, but necessary, due to the liberty that IP gives underlying 
network technologies to not use forward error correction bits.  If 
underlying technologies employs ECC, then this would be less of an, 
though certainly not a non-existant, issue.  For more information on 
this, see Phil Karn's documentation 
(http://www.ka9q.net/papers/newlinkpaper.ps.gz) for more details on what 
would be necessary to ameliorate the work-arounds that TCP implements to 
make up for the underlying protocol's lack of integrated support for 
reliability.  Note the paper is in the context of an amateur radio 
application, but the information presented is universal regardless of 
user domain.

--
Samuel A. Falvo II


---------------------------------------------------------------------
To unsubscribe, e-mail: colorforth-unsubscribe@xxxxxxxxxxxxxxxxxx
For additional commands, e-mail: colorforth-help@xxxxxxxxxxxxxxxxxx
Main web page - http://www.colorforth.com