[Cryptech Tech] Plans on the ARRGH project

[Benedikt Stockebrand]

Hi folks,

while I'm still struggling to adjust to a constant work overload, as
opposed to bursts of utter madness followed by times of recovery where I
could catch up with things, here's what I want to work on next year if
time permits---despite the fact that it already looks like the project
I'm currently in will continue to keep me rather busy throughout next
year:

  - I haven't gotten anywhere near coding yet, but I still want to write
    a test framework and some tests explicitly targeted at HWRNG output.
    It'll be most likely written in C++, not because I like that
    language but because I need the performance as well as the
    possibility to come up with some reasonable interface design.

    This is probably the most interesting plan to the Cryptech project,
    especially so with the various discussion on phenomena people have
    observed in their test data during the last few weeks.

  - I am going to split the ARRGH design into an analog noise source
    section, and a digital microcontroller/usb(?) interface section.
    The analog section should provide an alternative to Fredrik's
    boards, hopefully increasing the diversity in the Cryptech design,
    or at least providing a means for comparative testing.

  - While I'm at it, I want to have two different board designs, one
    using through-hole only and the other using SMD to go for more
    convenient size.

    This is where my personal goals differ somewhat from Cryptech: I
    want to come up with a design that even some slightly hamfisted
    enthusiasts can build themselves, as well as one that one might be
    more convenient but also somewhat challenging to build at home.

  - Once the digital part is separated from the analog, I'll see how to
    get rid of the FTDI chip.  While not immediately relevant to the
    Cryptech HSM, this might be of interest at least to Fredrik.

    Right now the most interesting approach seems to be using a
    PIC18F{13,14}K50 as some sort of not-so-drop-in replacement for the
    FTDI; it's a 20 pin MCU (available in through-hole as well as SMD)
    with both a serial and a USB interface.  Problem here is that the
    open source C compiler I found for this (sdcc) still seems to be
    work in progress, so this might be a dead end.  However, if it does,
    then this might help a lot of other people to get rid of the FT232
    in their designs as well.

    In that case the PIC18 doesn't work out I'll try the Microchip
    MCP2210 next; it's an SPI<->USB interface chip,and it's not
    available as through-hole (but at least as SOIC), but it's the
    second best candidate I know of right now.

    If that doesn't work out either, I guess I'll go for another, larger
    MCU with built-in USB support; maybe an Atmel ATxmega, maybe an
    STM32, or whatever.

    And yes, I know that this'll imply some driver programming on the
    host side as well...

And if things work over the holidays, I'll also finally do another video
on my blog talking about all this more extensively (and more to a not so
deeply involved audience).


Anyway, I'll try to find a way how to keep up with the mailing list more
closely next year.  See if I find one...


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/