# [RC5] Suggestion for when we Recycle Keys

Michael K. Weise mkw at att.net
Tue Jan 12 17:20:40 EST 1999

```Sorry, but your logic is a bit tangled.

Fact is that the probability of an unreturned block being returned within
the next X hours relates inversely (though not necessarily in linear
proportion) to the time that has passed since it was last issued. Thus, the
optimum order in which to reissue DES blocks is the one we're currently
using.

Michael

Paul Goyette wrote:
>
> Actually, if we started recycling somewhere in the middle, we'd get the
> best of both worlds.
>
> In other words, we start issuing keys at block 0x00000000, and we get to
> the last block issued, but only 16% have been returned.  Somewhere along
> the way, there'll be a point at which the density of returned blocks
> goes from > x% to something < x% (for some arbitrary x).  Say that that
> point occurs in block 0x01000000.  Then, we can start recycling keys at
> that point, or maybe at twice that point.  This way, we get to reissue
> key blocks as soon as we need to, but leave a substantial number of
> early-issue key blocks for the early folks to finish up their buffers.
>
> We could easily vary the x% based on the number of passes through the
> key space.  First pass set x% to 0 so we start at the very beginning of
> the key space.  Second pass, start at the point where less than 25% of
> the keys have been returned.  Third pass, start at the point where less
> than 50% have been returned, and so on.
>
> On Mon, 11 Jan 1999, Scott Dodson wrote:
>
> > The only problem with this solution is that it will allow less time for the
> > blocks that have been out there just a short time to actually be worked on,
> > it only makes sense to run through it once, then start again at the beginning
> > chronologically.
> >
> >
> > John McIntyre wrote:
> >
> > > I was also thinking about where to start recycling keys.  If the keys
> > > come back somewhat linearly, and we have issued 100% and checked 16%,
> > > does it make sense to start recycling at some percentage more than
> > > 16.1%?  Say half the difference between completed and the percentage
> > > over a hundred? For example Day 2 at 19.32% completed and 131.03% issued
> > > (131.03 - 100) - 19.32 = 11.71 /2 = 5.86 + 19.32 = 25.18 %
> > >
> > > This would allow the machine first issued that block a slighter longer
> > > chance to get it back while still allowing a faster machine a chance at
> > > blocks a little further through the key space, hopefully eliminating
> > > some duplication.
> > >
> > > --
> > > To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> > > rc5-digest subscribers replace rc5 with rc5-digest
> >
> >
> >
> >
> > --
> > To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> > rc5-digest subscribers replace rc5 with rc5-digest
> >
> >
>
> -----------------------------------------------------------------------------
> | Paul Goyette      | PGP DSS Key fingerprint:   | E-mail addresses:        |
> | Network Engineer  |   BCD7 5301 9513 58A6 0DBC |  paul at whooppee.com       |
> | and kernel hacker |   91EB ADB1 A280 3B79 9221 |  paul.goyette at ascend.com |
> -----------------------------------------------------------------------------
>
> --
> To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
> rc5-digest subscribers replace rc5 with rc5-digest

--
Michael                        [writing to you from 40°15'37"N 74°57'49"W]
http://home.att.net/~mkw
**************************************************************************
*       Michael K. Weise           Helping to get American products      *
* German Localization Services         ready for German-speaking markets *
**************************************************************************

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

```