Feeds

back to article Chinese IEEE members want MAC control for cognitive radio

For those in the mood for deep wireless communication geekery, a group of Chinese researchers has released a proposed protocol for MAC layer behaviour of cognitive radio systems. The idea is to increase spectrum efficiency by allowing wireless systems to be opportunistic in their use of spectrum: that is, to use whatever appears …

COMMENTS

This topic is closed for new posts.
Silver badge

Things that make you go "Hmmm".

Wouldn't this make truly "private" broadcasting nearly impossible?

I wonder why China would be interested in that?

0
0
Silver badge

Privacy is orthogonal here.

Re: "Wouldn't this make truly "private" broadcasting nearly impossible?"

Nope.

1
0
Silver badge

Re: Privacy is orthogonal here.

Nope. Not orthogonal. MAC exactly fingers particular hardware, by definition (NE1000/2000-clone cards from Taiwan in the late 1980s not withstanding). If you need a MAC address to broadcast, your broadcast is fingerprinted as being from your kit.

And yes, I know, I can fake a MAC address on this particular computer. Do you really think that the Chinese Radio technology will allow this prospect?

0
2
Silver badge
FAIL

Re: Privacy is orthogonal here.

And you think your mac isn't already being broadcast over the net? along with your IP address????

this is just moving away from fixed line and pre-defined wireless frequencies to unmanaged spectrum and more efficient usage...

And if your worried about MAC's identifying you, then change it each and every time on your broadcast node!

It is no different to having a Call Sign on HAM...

And it is not 'Chinese Radio technology' its is a proposed standard, multiple manufacturers would use this standard if it came around, and yes they probably would allow changing of MAC addresses... I know certain data centers will replace faulty hardware and reset the mac on the new NIC and even re-write the hardwares serial number when hardware is replaced... it would be very critical than MAC's can be changed...

0
2
Silver badge

@jake: MAC not MAC-Address

And no, they don't mean an Apple Macintosh either.

0
0
Silver badge

Re: @jake: MAC not MAC-Address

How, exactly, do you think a medium access controller works, Christian?

0
1
Silver badge
Thumb Up

Re: Privacy is orthogonal here.

"And you think your mac isn't already being broadcast over the net? along with your IP address????"

Not necessarily. The MAC address of the transmitter is going out, but that may not be your MAC address (if your transmission is being repeated, for example).

MAC addresses aren't usually passed beyond the first router. So if you're going through some kind of access point, the MAC that goes out is that of the access point, not the one of your computer. Of course, you may own the access point or router, so in that case, it would be "your" MAC address, but it would be lost (actually, replaced) as the packet goes through the cable company's router.

1
0
Anonymous Coward

Re: @jake: MAC not MAC-Address

@Jake - how do YOU think a media access controller works? News flash: MAC is a generic term; not all media has the concept of a unique address for a device, so not all media access controllers have a media access controller address. Just because all the media with which you are familiar has a MAC address does not mean that all media has a MAC address.

1
0
Silver badge

Re: Privacy is orthogonal here.

"MAC addresses aren't usually passed beyond the first router"

This used to be true in IPv4.

IN IPv4 the MAC is usually part of the IP address

MACs can (of course) be altered to something else via software, so fingerprinting isn't that useful. All that matters is that 2 identical MACs aren't on the same LAN.

0
0
Anonymous Coward

Re: Privacy is orthogonal here.

"This used to be true in IPv4.

IN IPv4 the MAC is usually part of the IP address"

1) I think you meant to say "In IPv6 the MAC is usually part of the IP address".

2) That is no longer true for the routable IPv6 address (the link-local address MAY, or may NOT, have the MAC as the lower octets), precisely due to the security issues.

0
0
Bronze badge

Re: Privacy is orthogonal here.

Would that be the same company that uses the MAC address as the serial ??

0
0
Silver badge

Hopefully China will continue

The logical endpoint of this is to realize that all of the available channels, with the possible exception of emergency spectrum should be repurposed and converged to carry TCP/IP (or a successor) network traffic.

With luck, China will pave the way so we can release ourselves from the stranglehold rent-seekers currently have on our bandwidth. As I write this, one of my bandwidth providers charges $0.20 per gigabyte for bandwidth while another is blithely charging around $1,000,000 per gigabyte. Same global network, just a different tarrif arbitrarily charged for a certain type of traffic because we let them get away with it. [The expensive one is based on charging for SMS messages]. On my local network, it is too cheap to meter.

I have not done the figures, but I am pretty sure we could afford to attach all of humanity to a global high-speed network and give everyone free access to all of the world's art, music, literature -- any cultural artifact that can be digitized. All we have to do is remove the rent seekers barring the gates.

0
1
Silver badge

Re: Hopefully China will continue

You really don't understand how radio works it seems. Two problems:

(A) There is simply not enough usable[1] spectrum in a typical built-up area to give everyone gigabit links like wired/fibre can.

(B) Cognitive radio is only usable if everyone is sharing the same negotiation system, and ideally (and essentially for non-cognitive systems) you have some database of users so you don't have a white-space user talking over another radio link it can't sense (due to sensitivity or shadowing issues) but is actually in use.

Hence radio should really be reserved for things that need it, such as mobile use, and not as a cost-saving option to avoid investing in fibre, etc, that has enough capacity for the foreseeable future.

[1] Yes, I know there is huge and very sparsely used spectrum up at tens or hundreds of GHz but that is not going to effectively penetrate well through walls, etc, and currently it is not cost-effective to produce TX/RX for them in single chips, etc.

4
0
Silver badge

@Paul

"(A) There is simply not enough usable[1] spectrum in a typical built-up area to give everyone gigabit links like wired/fibre can."

In German there is a word "Jain" to best reflect what I have to say to it. It's a combination of yes and no.

Hypothetically, if we use beamforming to a massive degree and lots of base stations, we might get to multi gigabit rates without bottlenecks. Unlike current idiotic passive optical networks, you can get around the shared medium problem with radio, at least in theory.

0
0
Silver badge

@Christian (was: Re: @Paul)

"In German there is a word "Jain" to best reflect what I have to say to it. It's a combination of yes and no."

I don't know enough Deutsch to grok "jain" ... perhaps "mu" is close in the internet's lingua franca?

0
0
Silver badge

Re: @Christian Berger

But is it practical to achieve that sort of tight beamforming in small enough packages and at reasonable cost, power consumption, and bandwidth?

If you consider the wavelengths for which we can probably use and get reasonable building penetration, say in the 10cm range, and an angular requirement of say 10 deg to get a large number of urban users then an ideal antenna needs a diameter of around 60cm and in practice a phased array is going to be significantly bigger. For a big base station that is viable, but not for handheld or set-top boxes.

For the OP's point about not going through service providers via white space mesh connection then it won't be practical, but for some future high rate mobile it might just be viable. If it is cost effective enough and has enough back-haul bandwidth.

0
0
Anonymous Coward

Beamforming

Even with beamforming (even with ANY directional antenna!) you have sidelobes of the signal - your main lobe may be pointing at you, but there WILL be sidelobes pointing elsewhere, causing interference. Yes, those sidelobes will be somewhat lower power, but still present.

0
0
Silver badge

Re: Hopefully China will continue

"(A) There is simply not enough usable[1] spectrum in a typical built-up area to give everyone gigabit links like wired/fibre can."

That is entirely dependent on how far the Gb link needs to be transported.

Before cellular technology was developed that same argument was raised against the possibility of widespread personal phones.

0
0
Anonymous Coward

it's likely going to enable 'pay-per-microsecond-per-herz'

and I might have to use my worldwide mesh-array of 'naked' (ultra-QRP) Raspberry-Pi's running WSPR HF beacon-mode operating below the noise level....if the prices aren't reasonable

2
0
Silver badge

Re: it's likely going to enable 'pay-per-microsecond-per-herz'

There would be a power factor in it, too. For example on CDMA networks you already have all mobile stations using the same band at the same time.

0
0
Silver badge

Ethernet....

"famous Ethernet model of “check, back off, wait a random amount of time, and check again”"

Shouldn't that be "send, back off if collision, then wait a random amount of time, and send again" ?

0
0
Silver badge

I know a little about radio ...

@Paul:

Re:You really don't understand how radio works it seems. Two problems:

(A) There is simply not enough usable[1] spectrum in a typical built-up area to give everyone gigabit links like wired/fibre can.

As @Allan intimated, we get around this by carving a given spectrum into 'cells'. The EM spectrum is only needed for the link to the nearest tower. The denser the population, the more towers. If it came to it, you could probably transmit terabits or more in a short line of site link without interfering with anyone else.

I was not suggesting we backhaul with EM spectrum. I don't think anyone else is seriously expecting the world's backbone to be anything other than something like fibre. The EM portion through the airwaves is just to get it to the physical network.

Re: (B) Cognitive radio is only usable if everyone is sharing the same negotiation system

The same system is already in place, we are using it now. It is called the Internet. If you are talking about negotiating protocols, as far as I know, the 'cognitive' of cognitive radio means the radio intelligently uses the best channels available to satisfy its needs. It only differs from WiFi in that uses more frequencies more intelligently, changing transmission/reception parameters -- so called 'dynamic spectrum management'. Otherwise, I expect it to use the same gateway and name server on-ramp to get on to the backbone and from there it is anybody's guess where it goes. Maybe we can coin a term 'meta cognitive radio' for one that is aware of even more protocols than you think they normally have.

I think it was Robert Ornstein who said something to this effect: "We should not be limited by what we believe is possible"

0
0
Silver badge

Lively conversation

For the record, I am very much concerned about issues of privacy and freedom of speech. I do not have any more or less faith that the Chinese may be up to no good than any other state. They all give me the willies.

The Chinese may or may not be ulitimately up to no good, but that does not mean that everything they do along the way is necessarily evil in itself.

Convergence of spectrum is a good thing because it increases available bandwidth. Even if we saturate the entire EM spectrum, we will *still* be starving for bandwidth. It is a fundamental limiting factor and any way around it is a 'good thing' (TM).

I expect that the Chinese, like any other powerful entity, would like to control the pipes or at least be able to snoop on what they are carrying. There are some ways to minimize the impact of that:

1) Encrypt everything

2) Make it dangerously illegal to spy on people without judicial oversight

3) Make the results of improper investigations entirely inadmissable as evidence

0
0
This topic is closed for new posts.