[RC5] DES 2 (in preparation of)

Dirk Kuypers dk at comnets.rwth-aachen.de
Wed Jan 7 09:39:23 EST 1998


On Tue, 6 Jan 1998, Martin Brandt wrote:

> >>Even with the current block size a pokey machine can get through it in
> >>less than a day.
> >
> >Yes, but then a lot of people might set up high numbers of buffers and only
> >flush them when they've done them all.  You could have an ancient machine
> with
> >a winning key hanging around trying to do another 50 blocks.
> 
> then why mess with the block size?  you'd have to increase the maximum block
> download to accomodate faster machines, and then this same "lot of people"
> could have their machine doing 5000 blocks before reporting..  with the
> current block size any machine *could* be setup to report daily.

BTW, is the buffer really searched in LIFO order? Then there could be
the problem of some blocks being buried by other blocks for a long
time on some machines. For example I have a buffer of 80 blocks at
home, my DX4 does about 30 a day and these 30 are normally flushed
daily. Even when I don't flush for some reason, I have 'fuel' for
another 1.5 days, so it happens very seldom that the complete buffer
is exhausted.

And the next (logical) question is, how do the perproxies handle their
buffers? The proxy I set up at the institute has 800 blocks in it,
but those are flushed very regulary. If the blocks are handled FIFO,
they are processed at the moment inbetween 8 hours, if it is LIFO, I
think I still have some blocks from November in there...

Dirk

--
To unsubcribe, send 'unsubscribe rc5' to majordomo at llamas.net
rc5-digest subscribers replace rc5 with rc5-digest



More information about the rc5 mailing list