LWN: Comments on "Regulating wireless devices"

Regulating wireless devices

jlokier — Wed, 27 Aug 2008 12:02:51 +0000

I wouldn't be surprised if the PDAs are told by their wi-fi base station that they are connected to an "indoor" device.

From a perspective of managing the radio spectrum, the restrictions themselves seem quite sensible to me. Everyone transmitting at the higher power outdoors would just raise the noise floor in densely populated areas while everyone's competing rather than cooperating. (And it would raise the interference between neighbours in houses too - already a problem with wi-fi in some areas, where you might detect 30 base stations from a single room.)

Far better to regulate in such a way that people have to respond by installing a finer mesh of base stations and developing more cooperative protocols (eventually...).

Regulating wireless devices

cortana — Sat, 23 Aug 2008 12:43:22 +0000

This is interesting. The manual for the wireless access point I purchased yesterday says the following:

> INTENDED USE. This device is a 2.4 GHz wireless LAN transceiver,
> intended for indoor home and office use in all EU and EFTA member states.
>
> ...
>
> POTENTIAL RESTRICTIVE USE. This device is a 2.4 GHz wireless LAN
> transceiver, intended for indoor home and office use in all EU and EFTA
> member states, except in France, Belgium and Italy where restrictive use
> applies.
>
> In Italy the end-user should apply for a license at the national
> spectrum authorities in order to obtain an authorization to use the
> device for setting up outdoor radio links.
>
> In Belgium there is a restriction in outdoor use. The frequency range in
> which outdoor operation in Belgium is permitted is 2460-2483.5 MHz,
>
> This device may not be used for setting up outdoor radio links in
> France. For more information see http://www.anfr.fr/ and/or
> http://www.art-telecom.fr

So it appears it's up to the end user to configure the device correctly in Belgium... but no hint as to the permitted channels is given, the user is expected to work them out from the frequency himself!

Regulating wireless devices

liljencrantz — Fri, 22 Aug 2008 09:39:55 +0000

I guess that the illusion of safety is more important than real safety here. If there is a
separate subsystem with cryptographically signed data about emission levels, that makes it
feel less open, less unrestricted, even if you in fact can just edit the source code a bit to
bypass the whole thing. Kind of like how the anti virus people want deep kernel hooks for
their anti virus products, even though there is nothing they can do with those that can't be
done by a pure user space solution. Kind of like how it has been repeatedly demonstrated that
there are many simple ways to get various weapons past airport security, but regular people
aren't even allowed to bring a bottle of shampoo.

Sometime security theater is the only thing that matters.

The good thing about this proposed subsystem is that it will actually make it easier to write
law abiding, bug free and stable drivers since it's suddenly much less work to find out at
what power you're allowed to transmit on a given frequency. So it serves a dual purpose of
both giving an illusion of safety and making the life of driver writers easier.

And hey, anything that will get this stupid policy deamon for my Intel wireless of my system
is a win in my book.

Regulating wireless devices

dlang — Thu, 21 Aug 2008 23:20:21 +0000

it's not the kernels responsibility to enforce the local law, but it is the users
responsibility to _comply_ with the local law.

and if the kernel doesn't provide reasonable tools to allow the user to comply with the local
laws then the kernel is at fault.

remember that if someone really wants to they can change the code (they have the source), and
that there are people who have legitimate reasons to operate outside the 'normal' rules (for
example Ham radio operators can use wifi equipment on channels that are not legal for other
people to use in the US)

Regulating wireless devices

sflintham — Thu, 21 Aug 2008 22:30:17 +0000

I am kind of replying to both previous comments here, so I appreciate I am not making fair
statements, but I'm just curious rather than trying to tell anyone they're wrong.

So in Belgium, people routinely make sure to tell their Pocket PDAs and similar gadgets they
are outside? And when they go back inside?

Or does no wireless device configured to know it's in Belgium ever use the lowest part of the
"a" domain at the higher power, just to be safe?

Or is this just a case of stupid legal distinctions which no one respects in practice?

Or - perhaps the most plausible guess - do portable devices never use that part of the domain
at the higher powered level, while mains powered and hence legally-presumably indoor devices
do?

Regulating wireless devices

nicolas@jungers — Thu, 21 Aug 2008 20:25:19 +0000

In Belgium the lowest part of the "a" domain (4 bands around 5.2 GHz) is restricted for indoor use at 200 mW and allowed for outdoor use at 25 mW. Values are from memory but the principle is still clear in my mind. A typical household external wall is an obstacle of around -10 db, sometimes more.

For info, an attenuation of 9 db lowers the signal from 200 mW to 25 mW.

Regulating wireless devices

mb — Thu, 21 Aug 2008 19:19:29 +0000

> And how does the device know it's outside?

The user may be forced by law to tell the device/driver.

Regulating wireless devices

sflintham — Thu, 21 Aug 2008 18:24:52 +0000

> There's also a set of flags for special rules; some domains, for example, do not allow
outdoor operation or certain types of modulation.

Just out of interest, where in the world is outdoor operation restricted? And how does the
device know it's outside? I suspect I've missed something really obvious here, but I have to
ask...

Regulating wireless devices

pr1268 — Thu, 21 Aug 2008 18:01:57 +0000

Keep in mind that if you do want a device to emit radio waves at frequencies and powers that are illegal in your current location, it is trivial to do so by simply using components that can be bought for a small number of dollars at the nearest radio shack and information that can be easily googled, or by making semi-trivial changes in a linux driver, or by slightly altering the hardware itself.

Despite the fact that this is common knowledge, it's still odd how the Wi-Fi vendors continue to "hide" behind the false perception of Linux users being a bunch of criminal hackers in order to justify not releasing hardware drivers. While there may be some substance to the notion that these companies are surrounded by market, legal, and trade-secret boundaries, I personally think that they're just too lazy to support anything other than Windows or Mac OS.

</vented frustration>

Regulating wireless devices

liljencrantz — Thu, 21 Aug 2008 09:38:46 +0000

Presumably, people in general don't actually _wan't_ to break the law. As such, a computer
subsystem that try to keep you from unintentionally breaking the law should be seen as a good
thing.

Keep in mind that if you do want a device to emit radio waves at frequencies and powers that
are illegal in your current location, it is trivial to do so by simply using components that
can be bought for a small number of dollars at the nearest radio shack and information that
can be easily googled, or by making semi-trivial changes in a linux driver, or by slightly
altering the hardware itself.

Regulating wireless devices

pabs — Thu, 21 Aug 2008 09:17:24 +0000

I like this development because;

It makes it easier to comply with local laws; arrest and fines are annoying, especially when
travelling.

Perhaps in future intel wireless will no longer need non-free binary-blob firmware:

http://bughost.org/bugzilla/show_bug.cgi?id=1594

People with ham-radio and other licenses may be able to experiment with software radio stuff
while easily complying with their special license.

Regulating wireless devices

dgm — Thu, 21 Aug 2008 08:25:58 +0000

Am I the only one that thinks that it's not the kernel's responsibility to enforce local law?
Should the Linux Kernel prevent being used in any other illegal actions? Maybe stop speeding
cars? Avoid being used in Internet fraud? Refuse to participate in a murder?

It is clear that the system should be able to impose limitations, but those should be decided
by the system administrator, not by the system itself. Most users cannot tell the difference
between a kernel enforced policy and a userspace (but privileged) one. What's more, most users
wouldn't event want to. And for those who would, it's trivial to build an emitting device.
This solution solves effectively nothing. The FCC should be equally served by an interface
that allows userspace to set the policy as required.

One final thought: laws and regulations change, not only from place to place, but with time
too.