You're right, there is. You describe it yourself:
> Congestion is not a temporary problem that can be fixed in such
> a simple manner. Even if every client does the backoff (hah!)
> there will be plenty of packets to which it does not apply.
>
... which is why the backbone routers now start doing the backoff algorithm
themselves: the more packets you send them without regard for proper
backoff algorithms, the more likely it is that _your_ packet will get
tossed.
> Perhaps a greedy algorithm would make more sense on the Internet.
>
Yep. It would kill the Internet. Really really fast.
This has already happened. In the "early" days, when people noticed that
bad backoff algorithms not only can, but _do_ kill networks. The heap of
very scientific papers Alan wrote of didn't get written in a vacuum, they
got written to explain, and fix, a very real problem.
Anyway, if you ever get to the point where the backoff exceeds more than a
few seconds you have an overloaded line. Jane User will kill her connection
once or twice, maybe, and then give up (or develop some patience), which is
no real problem. Your solution does neither, which is.
-- Matthias Urlichs | noris network GmbH | smurf@noris.de The quote was selected randomly. Really. | http://www.noris.de/~smurf/-- I'd love to, but I have to knit some dust bunnies for a charity bazaar.- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/faq.html