Re: encrypted cellphones

At 07:18 PM 1/30/96 -0500, prmoyer@magpage.com (Philip R. Moyer) wrote:
>Well, I'm discouraged.  I'm looking for strongly encrypted cellular telephones,
>but I can't seem to find many.  
Are you looking for cellphones, or cordless phones?  There aren't a lot of
strongly encrypted cordless phones out there, but there may be some.
Digital spread spectrum is probably the best that's easy to find;
other kinds of "digital" phones usually pick a not-too-busy frequency
and transmit digitized voice, which is mildly secure against other people
using your base unit to make their phone calls, but doesn't protect your
privacy against anyone with digital-capable equipment.

The middle ground between cordless phones and cellphones includes
cordless phones with ranges of about a mile (AT&T and some other vendors
have sold them); they're typically in the $300-500 range, and use
spread spectrum to avoid interference to/from other phones.
It also includes phone services that can handle portable phones that
you have to stay in one place to use (i.e. once you start your phone call,
if you go out of range your call gets dropped rather than handed off 
to another cell.)  I'm not aware of commercial service like this in the US,
but there are wireless PBXs that work this way (which can be cheaper than
stringing phone wires around buildings.)

Cellphones, of course, can only (usefully) use encryption if the
cellular service provider uses it (i.e. if the end that's listening
to your radio transmission can decode it :-)  American cell-phone 
providers don't.  The GSM phones used in much of the world have encryption,
but it's apparently not very strong.

>I would really like to avoid using a GAK enabled product,
>if there's any way to avoid it (even if it means paying lots of extra $$$).

I'm not aware of any GAKed cordless phones, though I supposed there could be
such.
US cellular phones don't need GAK because the government's strong-armed the
standards committees into using appallingly trivial crypto - none of this
strong 40-bit RC for you :-)
#--
#				Thanks;  Bill
# Bill Stewart, stewarts@ix.netcom.com, Pager/Voicemail 1-408-787-1281
# http://www.idiom.com/~wcs

On Wed, 31 Jan 1996, Bill Stewart wrote:

> providers don't.  The GSM phones used in much of the world have encryption,
> but it's apparently not very strong.

GSM A5 security is supposed to have effective key length of 40 bits, 
although according to some sources 64-bit session key is used. The 
algorithms are not freely available, so you never know.

I would say GSM security is still better than nothing. The problem is of 
course that only tha radio link is encrypted, not the connection out into 
public telephone network.

Juri Kaljundi, DigiMarket
jk@digit.ee

> I would say GSM security is still better than nothing. The problem is of 
> course that only tha radio link is encrypted, not the connection out into 
> public telephone network.

As I remember from discussions with a GSM encryption programmer
(which I posted to this list months ago) GSM is secure enough to prevent
real-time decryption, as keys are changed frequently using another
secure protocol (A8 I believe). GSM  encryption is only supposed
to make people like Princess Diana more secure, that's all. This was
enough for Pakistan to temporarily shut down Motorola's GSM network
in Karachi last February, until they discovered they could
intercept calls simpply by sitting at the base stations where
they're decrypted...

Rishab


> Juri Kaljundi, DigiMarket
> jk@digit.ee
>

Juri Kaljudi wrote: 

> 
> On Wed, 31 Jan 1996, Bill Stewart wrote:
> 
> > providers don't.  The GSM phones used in much of the world have encryption,
> > but it's apparently not very strong.

	As the A5 algorithm has so far not been publically disclosed, no
one outside of the spook community really knows if has a backdoor or
what computational effort might be involved in brute forcing it.  One
can certainly suppose that there was a lot of pressure to weaken it,
but whether that was accomplished by installing trapdoors or simply by making
special purpose hardware brute forcers simple, fast, and cheap is not known.

> 
> I would say GSM security is still better than nothing. The problem is of 
> course that only tha radio link is encrypted, not the connection out into 
> public telephone network.

	I have seen news stories about some shady "spy-shop" type
companies in England who are selling microwave receivers capable
of intercepting and decoding the microwave backhaul links that connect
most GSM cell sites to the mobile switching offices.   Apparently even
some supposedly secure GSM systems use unencrypted backhauls which can
be relatively easily intercepted by someone with the right gear 
from places near enough the towers to have a line of sight view
of them.

						Dave Emery
						die@die.com

> Cellphones, of course, can only (usefully) use encryption if the
> cellular service provider uses it (i.e. if the end that's listening
> to your radio transmission can decode it :-)  American cell-phone 
> providers don't.  The GSM phones used in much of the world have encryption,
> but it's apparently not very strong.

Don't forget the more attractive option: End-to-end.  Why leave the
plaintext available for the cellular provider?

Eric