[RC5] DES-II-3 [LONG] [Repost]

dan motion at es.co.nz
Wed Sep 2 22:25:40 EDT 1998


On Tue, 1 Sep 1998 16:16:33 +0100, David Taylor wrote:

>Increasing the proxy network could work, but there would still be a very
>large load on them if every client working suddenly connected at the same
>time

Yeah, but how likely is it that every client is going to connect at once?

Based on calculations i posted to the list just after DES-II-2 the current sized proxy network should be able to handle a scheme where the clients are 
programmed to connect at the very start of DES-II-3 and get the new cyhper text.

Lets say , very approximatley 40,000 active machines (from SPEC/FLOPS page)
/20 proxies = 2000 machines per proxy
less 10% behind pproxies = 1900
less 20% offline and dialup machines =1700
now, every machine is not going to be synchronised, the times are going to be out by say +/-30 secs.
1700/60 secs = 28 /sec
Now the cypher text should be able to be transmitted in a second, given that it is dished out independantly from the blocks. This means an individual proxy 
machine needs only support 28 connections per second. easy peasy.

This assumes the clients are going to stop working on the current block and connenct the moment the alarm goes off. If rather, they checked the inbuilt 
alarm clock only at the finishing of a block, then the load would be spread much much further meaning my 486 OS/2 server could probably handle it all by 
itself.



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



More information about the rc5 mailing list