[RC5] Time out and DES-II-2 challenge

Joe Zbiciak j-zbiciak1 at ti.com
Wed Jul 1 10:09:32 EDT 1998


'Steve Bird' said previously:
| 
| If the key could easily be found using the 2^28 keysize then why even
| use the 2^30??  Could someone explain the significance of the size of
| the exponent (other than ....the bigger the exponent, the bigger the
| number)

2^28 refers to the number of keys in a block, not the size of the key in
bits. 

Here's my reasoning behind suggesting 2^28th blocks instead of 2^30th 
blocks for mostly-offline machines:

When you set your client to buffer blocks larger than 2^28th, it
requests the larger block size from the keyserver.  The keyserver then
looks for blocks equal to or smaller than the size you've requested
within its alotment.  If it can't find any, it carves a block of the
requested size out of a larger block and give that to you.

What this means is that if you're client is set to process 2^30 size
blocks, then you could receive 2^28, 2^29, 3*2^28, or 2^30 blocks. 
If you have an online client, this is fine.  A problem arises, though,
when you're trying to set up buffering for an offline client.  

A typical scenario for an offline client is one which can fetch and
flush at fixed intervals (eg. through cron jobs, sneaker net, whatever).
This sort of client needs a buffer size which matches the predicted
time between buffer updates.  With a 2^30 preferred block-size setting,
there is no guarantee on how long a given block will take to process --
there's a 4:1 difference between the largest and smallest block you 
might buffer.  

So, for these clients which need a predictable workload, it only makes
sense to buffer the smallest available blocksize, so that it is possible
to exactly control the amount of buffered work by setting the size of
the buff-in file.

Note that this could be fixed by changing the buff-in specification to
be the number of 2^28 equivalent blocks, so a 2^30 block would count as
four blocks in buff-in instead of one.  If we did that, then I would
change my dialup/offline clients to use 2^30 or 2^31-sized blocks
instead of 2^28 blocks.

Regards,

--Joe


-- 
  +------- Joseph Zbiciak ------+
  |- - - j-zbiciak1 at ti.com - - -|  "The truth of a proposition has nothing
  | -Texas Instruments, Dallas- |   to do with its credibility.
  |- - #include <disclaim.h> - -|   And vice versa."       -- Lazarus Long
  +-----------------------------+

--
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