[Cryptech Tech] Back to Cryptech, and some minor news

[Benedikt Stockebrand]

Hi folks,

after another couple of weeks of assorted madness going on around me
I've finally managed to catch up on the list, and by end of next week
there might be a chance that I'll find time to work on entropy sources
again.  So, to catch up with things:


@Bill: Would you like me to send you one of the ARRGH boards (entropy
       sources) I've got?  If so, just send me your snail mail address
       and I'll get it shipped to you.

@Randy: Sorry I missed your presentation at RIPE; somebody intercepted
        me right before, and I only managed to watch the video when I
        was back at Heathrow.  Anyway, Zener/avalanche diodes are
        specifically designed to take the abuse, so that's why I was
        reluctant to abuse BE junctions on BJTs, which *are* abused when
        using them like that.  And BTW, where have you been all the
        time?  I've tried several times to find you, but you always put
        on your invisibility cloak:-)

@Joachim: Have you tried to hook up the analog side of the ARRGH board
          yet?  I'm really curious if it worked for you, and if so, how
          the performance difference between Fredrik's board and mine
          was without a MCU in-between.


Currently my plan on the hardware side is to split the current prototype
into an analog noise source module and a digital processing module
again.  The results should be an SMD based analog noise source board
that can be fed with 3V3+ or whatever is available but otherwise doesn't
really differ from what the current ARRGH board has, and at least one
separate MCU/USB side board, partly with the goal to get rid of some
certain FTDI chip in the design.  I've spent a bit of research time
(mostly while sitting in a train) looking for viable options there, and
while none looked like what I'd really wanted, there are some I'm
willing to give a go.

The result of two separate subsystems would be that the analog side
could be more easily fit to the Cryptech boards and that way add
diversity here.  I also consider hooking it up straight to the GPIO pins
of a Raspberry Pi or BeagleBone Black to see how much output can be
extracted with something faster than the 20MHz MCUs.  And yes, this will
make testing significantly more difficult due to the timing constraints
of the operating system involved.  And no, so far I've never had reason
to gather experience writing OS drivers.  Looks like another fun new
experience.

And no promises, but I'll try to go to the CCC congress in Hamburg after
christmas and get myself synchronized with Fredrik there; let's just
hope nothing unexpected keeps me from going there...


The other end I want to work on is the test software side; I have some
ideas on writing a test framework that simplifies implementing whatever
tests anyone can come up with.  I'm currently considering two
alternatives, one providing an array-like interface with higher-level
filtering and manipulation operations, and another doing similar things
on a bitstream; both approaches have their advantages and disadvantes,
so I'll probably do some more thinking about it before I actually start
coding.


One more thing: Yesterday at the DENOG meeting in Darmstadt I've done a
presentation on crypto hardware, with particular references to both
Cryptech and ARRGH, and considering the questions following a lot of
people are really interested in our work.  

Before me, someone from Rohde&Schwarz who talked about their closed
design HSMs took so much heat I admit I felt kind of sorry for him;
apparently he had never encountered people with our kind of mindset...


Cheers,

    Benedikt

-- 
Benedikt Stockebrand,                   Stepladder IT Training+Consulting
Dipl.-Inform.                           http://www.stepladder-it.com/

          Business Grade IPv6 --- Consulting, Training, Projects

BIVBlog---Benedikt's IT Video Blog: http://www.stepladder-it.com/bivblog/