[RC5] Number of Participants Stats

gindrup at okway.okstate.edu gindrup at okway.okstate.edu
Thu Nov 20 19:05:53 EST 1997


     Extend the serial numbers by the MAC address.  Repeated MAC addresses 
     would be detectable and thus so would reloading the client.  However, 
     this does nothing for the machines that have no network card.  
     Microsoft uses a broken version of an International Standard to try to 
     hand out unique class instance numbers to OLE containers and their 
     method even works on NIC-less machines.  It is still *possible* to 
     spoof their method, just very unlikely.
     
     Both your method and the one your replying to attempt to measure the 
     number of machines running the client, which is not the same as the 
     number of people running the client.  My personal splay factor is 
     currently ~3 and will become much larger if a faceless Mac client is 
     released anytime soon.  I vaguely recall that the original poster 
     wanted to know how many *people* were participating.  Counting e-Mail 
     addresses is about the only way I can think to approximate this 
     number.
     
     To address your concern about theory...  If one defines the "average 
     machine keyrate" as the total effort keyrate divided by the number of 
     machines, then the method that you are attacking will actually work.  
     Surely, the more powerful machines will count too many times and the 
     less powerful machines will count too few.  But the definition of the 
     "average machine keyrate" is such that this exactly balances.
            -- Eric Gindrup ! gindrup at Okway.okstate.edu


______________________________ Reply Separator _________________________________
Subject: Re: [RC5] Number of Participants Stats 
Author:  <rc5 at llamas.net > at SMTP
Date:    1997/11/19 16:48


     
In the end, because (I believe) there are more slower machines than faster 
machines out there, you method would under estimate the total number of 
machines involved.  Not by several orders of magnitude, mind you, but by a 
significant percentage.
     
The only true way to measure haw many machines are participating is by 
handing out serial numbers.  The first time a client is run, the keyserver 
hands it a unique serial number that the client subsequently returns 
attached to each block.  Of course, this wouldn't solve the problem of 
someone running the client, deleting it, and then starting up again with a 
new client.  This method would OVER estimate the number of machines.
     
Well, I've come to the conclusion there isn't a way to solve this problem. 
Time to drown my troubles in some coffee...
     
Nathan Taylor
EMRTC Publications
     
--
To unsubcribe, send 'unsubscribe rc5' to majordomo at llamas.net 
rc5-digest subscribers replace rc5 with rc5-digest
     


--
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