RF> The Miss Manners approach to network programming dictates, however,
RF> that a single application not go about hogging resources (like
RF> sockets, bandwidth, etc...) to accomplish this.
KA> Commenting in response to the folks who say that a server reaches
KA> steady state whether the clients are single-connection or
KA> parallal-connection: I agree with the note that with faster user
KA> response, the user may read more pages over a given period of time.
KA> This will increase the steady state load by some amount. (My feeling
KA> is that it would be less than 25% for a population of serious
KA> browsers, and a higher percentage for a population of wild surfers)
I have a hard time complaining that Web usability has gotten more efficient
as a result of Mozilla multi-threading. Hell, if I wanted an information
system that was a pain for the users to use, I would've stuck with gopher :)
I am of the opinion that anyone wishing to claim that bursty traffic is
harder on the network as a whole would do extremely well to present some
amount of evidence, since such a claim tends to contradict conventional
wisdom in operating systems (queuing theory and the such). If I'm not
mistaken, the average waiting time should not increase, which is typically
the measure by which to judge network performance. This analysis of course
neglects Karl's observation that average usage may increase due to ability
to begin the next transaction before the current is finished. I addressed
this concern in the above paragraph.
-- Jared_Rhine@hmc.edu | Harvey Mudd College | http://www.hmc.edu/~jared/home.html"To hear many religious people talk, one would think God created the torso, head, legs and arms, but the devil slapped on the genitals." -- Don Schrader