-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Matthew Toseland wrote:
> I mean we won't get enough information to have a useful and accurate
> rate for each pair.
I see your point, but hopefully it won't be necessary to have a very
accurate rate for pairs that don't communicate much. The rate only needs
to adapt quickly if there's a large amount of traffic, doesn't it?
> This is not a serious problem, because we are not really dealing with a
> point to point routed network here. The same proportion of requests from E
> and from A will reach F. If D is overloaded, then everyone needs to slow
> down, because *everyone* is talking to D. If they are further away they
> probably talk to D less and so need to slow down less.
But that doesn't solve the problem of pushing the load back to A. If A
causes D to become overloaded and everyone else who's using D slows
down, A has no incentive to slow down. By pushing the load back along
the path of the request that triggered the overload, you can quench
sources that trigger more than their fair share of overloads.
But then the problem is how to avoid throttling other paths that happen
to share links with the path you want to throttle.
I understand that no path will necessarily be used more than once, but
if the throttlng travels back towards the source then that doesn't
matter: like radiation therapy, the various rate reductions will meet
near the source of the traffic.
> A-B-C-D-E-A (benzene ring)
>
> Congestion at E, so E-D slows down, C-D slows down, B-C slows down, A-B
> slows down, E-A slows down...
So the speed of every node in the ring is limited by the speed of the
slowest node... is this still true if the ring's embedded in a larger
network?
F---E
/ \
A D---G
\ /
B---C
\
H
If you only keep one rate per neighbour then congestion at E will slow
down CD, affecting HCDG. But if you keep one rate per pair of neighbours
then HCDG will be unaffected, because C will reduce its rate from B to D
but not its rate from H to D.
Cheers,
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2.2 (GNU/Linux)
iD4DBQFEPn7eyua14OQlJ3sRAj8bAKDkQI6wxN/b155dbgaZ670ggYyaIQCWNNd5
SzhGIKC/L1vWkpKeCuj1rA==
=yGNm
-----END PGP SIGNATURE-----
_______________________________________________
Devl mailing list
Devl@freenetproject.org
http://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
