[Tsvwg] algorithm 5 (was Re: WGLC for Port Randomization starts now (April 1st))

[Mark Allman <[e-mail]>]

> >   - 3.3.5: The algorithm is wrong.  This is my fault.  The paper is not
> >     clear.  Alas.  When a collision is found we do not increment by one,
> >     but choose a new random increment.  This is just not discussed in
> >     the paper, but should have been.  However, I looked at the code for
> >     my simulator and that is how I did it.
> 
> Any specific reason for this approach instead of the one described ni
> the port randomization I-D?

Let's backup and make sure we make something clear... I didn't somehow
mis-implement A5 in my tests.  The I-D is wrong.  (It is my fault the
I-D is wrong.  The paper is not clear enough.)  So, I'd prefer if these
were synced up just to avoid confusion.  If you want to do a riff, call
it something else.

> My take is that the one in our I-D could be better in terms of
> collisions of connection-id's.

I actually don't think it matters.  Look at the difference between A1
and A2.  Basically, there isn't any difference in the collision rate.
Sometimes A1 is better, sometimes A2 is better.  I'd expect the same
thing in the case of A5 and some inc-by-one riff.

allman