michael at memra.com
Tue Aug 6 03:15:27 UTC 1996
On Mon, 5 Aug 1996, Paul Ferguson wrote:
> Sorry to name specific products here (lots of folks specifically point
> out cisco, so what the hell), but in the specific case of Pointcast,
> there *is* a knob which controls the time-frame between updates. This
> is, in my opinion, the Right Thing to do.
How would you feel if Pointcast slowly backed off and increased the time
period between updates, but at the touch of a key or movement of the mouse
it would go back to the "knob" setting? Obviously a low volume data feed
like Pointcast may not be that big a load on the net but the heuristics
(which do involve human factors) could probably be applied to a lot of
other things like video feeds that will be bigger bandwidth consumers.
Of course, maybe we could have some sort of streaming feed discovery
protocol whereby a server can detect that multiple clients are receiving
the same stream and ask them to negotiate with each other and appoint a
master client that relays the feed locally to the other clients.
> Perhaps an FYI RFC on application behavior is not such a Bad Thing, but
> it would certainly be a controversial document.
This can be a good thing because it gets the issue noticed and some of the
people on the sidelines go off and do real work that can result in
resolving the real problems.
Michael Dillon - ISP & Internet Consulting
Memra Software Inc. - Fax: +1-604-546-3049
http://www.memra.com - E-mail: michael at memra.com
More information about the NANOG