Thermal Imaging Decision Applicable to TEMPEST?

John Young jya at pipeline.com
Wed Jun 13 07:36:22 PDT 2001


David Honig wrote:

>Two words: antenna design.

A third is signal analysis.

A principle argument against being able to sort through the geometric
increase in devices that leak emissions since the 1960s is that it is 
nearly impossible to find a pin in the hugely noisy haystack of the
electrogmagnetic spectrum.

Help me out here with signal analysis capability even with the niagara
of the digital age. Is it not possible to sort through a very large range
of signal using readily available algorithms to then pinpoint the
signature of types of sources, then home in on subsets of those
sources, to finally single out a particular source?

With the increase in signal volume has come a corresponding increase 
in signal analysis capability. Analysis of the full electromagnetic spectrum
has been possible for quite a while, if public documents on military research
are a reliable guide, and comprehensive analysis is ever being refined with
with increasingly fine granularity.

While there are billions of electronic devices leaking emissions, there
are no where near as many EM slots used by those devices and their
emissions. In fact, there are only a small number of public slots -- so 
long as devices conform to regulations. EM leakage is regulated as well.

If the world's devices conform to regulations, and those EM slots are
known and catalogued for signal analysis, then there is a question
about the leakage of the leakage, that is, emissions that escape
regulation, by poor device design, by granularity, or unintentionally.

The signature of a device which leaks, or makes noise, in a unique way 
is what presumably is searched for in sophisticated signal analysis. 
A few hundred submarines are identified this way, as are potentially
billions of people.

Are there too many unique device signatures to acquire and identify?
Perhaps so, but I suspect that enterprise is being diligently worked
on, beginning with data provided by manufacturers, catalogung
implanted emissive attributes in the devices, using benchmarks
for types of devices, tracking taggants and moles, cooking up
new variants on Hidden Markov and the host of search/sort/analyze/ID
algos.

Jumping off the cliff of ignorance, I suspect that signal analysis,
as with cryptanalysis, will be always able to find a way to get 
around obscurity. If you don't want to be acquired, don't signal.
Silencio, mafia.





More information about the cypherpunks-legacy mailing list