[2018] in cryptography@c2.net mail archive
Re: Mobile phones used as trackers
daemon@ATHENA.MIT.EDU (Phil Karn)
Tue Dec 30 17:44:28 1997
Date: Tue, 30 Dec 1997 14:36:22 -0800 (PST)
From: Phil Karn <karn@qualcomm.com>
To: davidw@datamgmt.com
CC: cryptography@c2.net, perry@piermont.com, karn@qualcomm.com
In-reply-to: <34A94699.25CDF3C2@datamgmt.com> (message from David M Walker on
Tue, 30 Dec 1997 19:08:09 +0000)
The article almost certainly refers to what the industry calls
"registration messages". These let the cellular network know which
cell you're in so incoming calls can be delivered directly to your
cell.
Cellular registration works much like bridging in an Ethernet network.
If the mobile has been heard from recently, the network can direct
that mobile's pages (incoming call notifications) to the cell in which
the mobile was last heard. Otherwise, the network can "flood page" the
mobile in all cells in the system.
Registration usually occurs when a cell site invites the mobiles in
its area to register themselves. This is an "explicit" registration.
This can occur automatically and completely without warning at any
time the phone is turned on. (Some phones, such as my Motorola Micro
TAC Lite, emit a soft click from the receiver when they register due
to mild RF interference from the transmitter to the audio circuits,
but this is clearly not a design feature.)
Implicit registration occurs when you make a call, much as it happens
when you send a packet in a bridged Ethernet network.
Registration is technically a carrier option. And it's a tradeoff
between decreased paging traffic and the overhead of the registration
traffic itself. But as cellular networks get larger, individual cells
get smaller and call traffic increases, the improved efficiency
becomes quite compelling. Intrasystem registration now seems pretty
much universal, at least in the US. My AMPS carrier here in SD, GTE,
requires a registration before it will even try to deliver a call --
i.e., there seems to be no flood paging at all.
Registration is *mandatory* when intER-carrier roaming is involved.
Some comments on the privacy implications of registration:
1. In my opinion, cellular registration is one of the most
problematical privacy issues in modern telecommunications. Unlike the
actual contents of a call, which can at least in theory (if not in
practice) be end-to-end encrypted against interception in the network,
registration information is user-to-network. The network needs that
information to do its job efficiently. I don't see how cryptography
can help here.
2. There is, however, no justification for *logging* registration
information. When the network wants to deliver a call to you, it needs
to know where you are *now* -- not where you were an hour or a day
ago. This seems like a good point on which to make policy.
3. Registration does actually provide one minor privacy enhancement
over flood paging. With flood paging, an RF eavesdropper anywhere in
the system can build a complete log of your incoming calls. With
registration, he has to be in the same cell with you.
The only countermeasure I can think of against registration tracking
is to keep your cell phone turned off anywhere you don't want the
world to know you've visited. One alternate way to receive your calls
is to carry a (one way!) pager. When you get a page, you then have the
option of turning on your cell phone (and revealing your location), or
returning the call on a pay or conventional telephone. Even the latter
technique runs the risk of having telephone call detail records
cross-correlated against a log of your pager messages, since the
latter are invariably in the clear.
Phil
