[109504] in Cypherpunks
Re: About Alpha radiation...
daemon@ATHENA.MIT.EDU (Jean-Francois Avon)
Fri Mar 26 19:35:12 1999
From: "Jean-Francois Avon" <jf_avon@citenet.net>
To: "Cypherpunks" <cypherpunks@toad.com>,
"Marty Levy" <rwww60@email.sps.mot.com>
Date: Fri, 26 Mar 1999 12:00:49 -0500
Reply-To: "Jean-Francois Avon" <jf_avon@citenet.net>
On Fri, 26 Mar 1999 07:31:52 +0000, Marty Levy wrote:
>Alpha particles are a major source of bit disturb errors in terrestrial-usage
>semiconductors. The primary source is the packaging used to encapsulate the
>chip. Low alpha mold compounds can reduce this in plastic encapsulated
>devices, but cannot eliminate it. The presence of alpha generators in
ceramic
>packages is even more difficult to control. Trench-style DRAM cells are less
>susceptible to alpha upsets due to the protection afforded by the silicon
>substrate. Multiple levels of interconnect also improve the situation, but
of
>course shrinking device geometries make it worse.
>
>There is a great deal of literature on this subject for anyone who really
wants
>to know. I would start with IEDM, IRPS (International Reliability Physics
>Symposium) and various IEEE societies dealing with solid-state devices.
Thanks for the info.
My original comment was aimed at that post:
----------
No,
my table companion worked for a private company and the government
in question is UK's. I asked him why government ciphers are still
implemented in hardware and not in software working in dedicated
one chip computers. Apparently speed is not the answer - very high
security communication is really low bandwidth. I didn't find his
answers very convincing: He said that an alpha particle from space
can knock out a transistor in the chip and modify the cipher. Then
again it is a trivial matter to have not one but many CPUs
----------
Ciao
jfa
