Networks and responsiveness (was Re: Marketplace story on FCC and Comcast {for sure djf})

[David Farber <dave () farber net>]

________________________________________
From: peter.capek () gmail com [peter.capek () gmail com] On Behalf Of Peter Capek [capek () ieee org]
Sent: Monday, July 14, 2008 2:19 PM
To: David Farber
Subject: Networks and responsiveness (was Re: [IP] Marketplace story on FCC and Comcast {for sure djf})

On Mon, Jul 14, 2008 at 11:35 AM, Tony Lauck <tlauck () madriver com> wrote:

Networks need to be designed so that the load placed by heavy users
doesn't unduly impact the responsiveness seen by light users. This is an
engineering problem that has been recognized since the design of
time-sharing scheduling algorithms in the 1960's.

This is correct, but perhaps too simplistic.  We should strive to find a way so that
interactive use of the Internet - ideally whether light or heavy - is not unduly affected
by non-interactive use of it.  With the rapidly growing trend toward hosted applications,
I am fearful that we'll go, once again, through a period of forcing users to experience
slowdown, with all the frustration, loss of attentiveness, and anger that it causes, as
many users did with overloaded mainframes in an earlier era.

If I'm using the network to do, say, a backup, and it takes 5 or 10 or even 50% longer because
some other user is interacting heavily with a network server and getting good responsiveness,
that seems to me to be a good trade-off, if I'll be similarly benefited when I'm engaged in such
use.  I've proposed a goal here; unfortunately, it probably can't be achieved in a way that isn't
subject to abuse, bits being the application-independent things they are.  But we should move
in this direction.

           Peter Capek


-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com