Are chip makers building electronic trapdoors in key military hardware? The Pentagon is making its biggest effort yet to find out
|
|||||||||||||||||||
|
| |||||||||||||||||||
|
« Taiwanese software spots stock-market stinkers | Main | The Mash Monsters »
Are chip makers building electronic trapdoors in key military hardware? The Pentagon is making its biggest effort yet to find out
TrackBack URL for this entry:
http://blogs.spectrum.ieee.org/cgi-bin/mt/mt-t.fcgi/4585
This page contains a single entry from the blog posted on May 1, 2008 4:16 PM.
The previous post in this blog was Taiwanese software spots stock-market stinkers.
The next post in this blog is The Mash Monsters.
Many more can be found on the main index page or by looking through the archives.
Comments (6)
Just like open source software and a thriving community is critical for security, same is needed with hardware.
FACT: The currrent culture in USA of lowest dollar wins, lawsuits, dual monopoly breadline only, is a massive threat to the infrastructure of the USA/world.
Until the NSA/DOD/Darpa/NSF/etc write a report on how LAME, crass, stupid their policies have been in the past, not much will change fast enough.
The history of UNIX is LAME, lawyers screwed up the business, a$$wholes Monopolized IT, and Everywhere, was junk. This allowed Mico$oft to grow. The same trend almost happened with the SCO lawsuit.
The leadership of the USA is Horrible! If we are a nation at war, we better start making sure we got the policy right. The leaders of this country, since the cold war ended, have really made serious OMISSIONS in important policy.
Write the Reports that RAND never wrote, or others.
It takes a civilian industry to defend the country. Basic principles. Lock things up, like Freedom of speech, gun rights, and innovation and freedom, everytime in history, Kaput!
Would you trust Microsoft or Intel with your country with how they do business? NOPE. Competition should be allowed and enforced to happen! FAILURE there!
ETC...I could write the report, but .....
Posted by 2CompetitionActivist | May 1, 2008 6:00 PM
Posted on May 1, 2008 18:00
When are the American manufacturers planning to take back our Country from China? It is high time that America produces sensitive military devices, not a Communist country that repeatedly produces tainted toys and food.
Posted by Carey Ann Ginos | May 1, 2008 6:22 PM
Posted on May 1, 2008 18:22
I highly doubt there were, "backdoors" in the hardware side of these semiconductors. It was more likely that there are software backdoors either accidental (buffer overflows, etc) or intentially.
For an example of a hardware based backdoor, my company found one produced by Atmel in their CryptoMemory family (http://www.flylogic.net/blog/?p=25).
More than likely, someone on the inside sold the secrets allowing an attacker to then enter into the system and disable it or lower it's sensitivity level for detection.
Nice article though!
Posted by Christopher Tarnovsky | May 2, 2008 5:01 PM
Posted on May 2, 2008 17:01
Your statement that all you can do is peel the chip and look for insertions is incorrect.
And it is not enough.
1, This will only find something like an incindiary destruct device or radio.
2, A chip has a limited number of inputs, or leads. You can exhaust the combinations of inputs and examine the outputs. This is essentially testing the definition of a particular chip's repitoir.
3, A more clever destruct would be performed with the use of only a single transistor on the circuit board, and floating gate to hold a value. The floating gate when activated holds the transistor open for the next ten years or so, the transistor brings a number of vital points in the chips circuits to ground. Killing the total functionality of the chip.
This method could be used to generate specific undesireable ouputs.
Only exhausting the second point would overcome the third. And you need to watch out for inputs dependant upon all possible states of the chip prior to the input, since it is possible to put a chip into a condition to perform a unique function based on an input which would not have the same effect under a different chip condition.
Posted by cyberbian | May 5, 2008 3:58 AM
Posted on May 5, 2008 03:58
I remember a curious little story related to me after a few beers by a guy who ran Harris mainframes for a local bureau. Early '80's. He claimed that, after an abend and a hard re-boot, the mainframe came back up with a screen indicating that it was 'expecting a DoD satellite' initiation sequence.
I had always assumed it was the beer. Maybe it really was the chips?
Posted by Waymad | May 12, 2008 1:28 AM
Posted on May 12, 2008 01:28
Perhaps it is time to produce these chips in the US. The cost of verification of foreign chips is probably higher than the cost of building them here in the first place...
Posted by G$ | May 13, 2008 7:26 PM
Posted on May 13, 2008 19:26