[rc5] RC64 keyspace

Eric Gindrup gindrup at okway.okstate.edu
Tue Oct 28 11:28:54 EST 1997


        I'm more worried about the storage method for assigned and returned 
     blocks.  There appears to be a difficulty where an interrupted client 
     loses a block.  By itself, this is irrelevant.  However, having single 
     blocks drop out here and there makes the set of assigned and not 
     returned blocks "interesting", i.e. hard to represent compactly.
     
        Assuming that one bit is used to record the assignment of a block 
     and one bit is used to record the receipt of a block, 2^65 bits will 
     be needed.  This is 2^32GB (4 Exabytes).  I don't know where *I* would 
     go to look for a storage system of that magnitude...
        If the set of "tardy" blocks was simple, then extent methods could 
     be used to represent the sets of assigned blocks and returned blocks.  
     Given the method of assigning, such a representation is probably 
     already in use for assigning.  This still leaves a potential for 2EB 
     of returned block data.  Of course, if the loss rate is < ~1/16, then 
     extents might be more efficient.
     
        In any event, I'd expect that the Bovine crew is going to need 
     about 256-times as much storage for this information as was used in 
     RC5-56.
            -- Eric Gindrup ! gindrup at okway.okstate.edu


______________________________ Reply Separator _________________________________
Subject: [rc5] RC64 keyspace 
Author:  <rc5 at llamas.net > at SMTP
Date:    10/27/97 10:46 AM


     
  I was really hoping that the D.N. wasn't going to move onto the 64bit 
challenge, but since we cracked the 56 so quickly I don't have anything 
else to use my cycles on.
     
 What I was really wondering was whether we could use some analysis 
techniques to check high probability key space segments first, instead of 
just randomly checking the entire keyspace?  I read some information on 
this early in the 56bit days, but don't remember the specifics.  Has this 
already been ruled out?
     
  -Mike Lingk
     
----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the
body.
     


----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list