Array
(
    [content] => 
    [params] => Array
        (
            [0] => /forum/index.php?profile-posts/145/load-previous&before=1702416050
        )

    [addOns] => Array
        (
            [DL6/MLTP] => 13
            [Hampel/TimeZoneDebug] => 1000070
            [SV/ChangePostDate] => 2010200
            [SemiWiki/Newsletter] => 1000010
            [SemiWiki/WPMenu] => 1000010
            [SemiWiki/XPressExtend] => 1000010
            [ThemeHouse/XLink] => 1000970
            [ThemeHouse/XPress] => 1010570
            [XF] => 2021370
            [XFI] => 1050270
        )

    [wordpress] => /var/www/html
)

SemiWiki

C
cliff
Are you talking about hiding the circuit? We are going to have to add bogus blocks for obfuscation, and add a decoder to burn out the chips unless certain pins are shorted to ground.

I will call up the mask making companies. I have an idea. I can explain privately.
T
Tanj
Reminds me of the story of the 16kb DRAM, where they embedded half a dozen traps in the design where a reverse engineer would see a standard element but not see a lower level feature which disabled it. That practice was used in earlier chips too, and became rife to combat designs getting ripped off. So, reverse engineers would have bugs that were subtle but product killers..
T
Tanj
I was just agreeing with you that reverse engineering from GDS-II or OASIS is probably possible, and would be easier than ever if the pirates are smart enough to use AI. A modern glitch might be something like a cell with a high recognition factor like an xor or a latch where a subtle change in geometry leaves it disconnected in fact, but where the AI might simply "match" it and move on.
C
cliff
Analog is our valued IP. We make SerDes, PLLs (also within SerDes), ADCs, etc. When we are making a custom ASIC for a customer, we are going with US citizen employees only. We are pushing the security angle. I am thinking about sending half the layers to 2 different mask shops. Have you ever heard of doing that?
Back
Top