Broadcom Accuses Atheros Of WiFi Pollution

eggboard writes "We just posted a story at PC World about 802.11g chipmaker Broadcom's claims that the high-speed 108 Mbps mode available in rival Atheros's AR5004G chipset disrupts all nearby Wi-Fi networks. The Turbo mode, part of Atheros Super G, uses two Wi-Fi channels (5 and 6) to double bandwidth, but Broadcom says this can lead to 'an enormous degradation in the speed of nearby 802.11b and 802.11g networks.'. D-Link and NetGear are shipping Super G-based devices. If Broadcom is right, Atheros gear would pollute neighbors' networks. If wrong, they're putting out a pretty heavy marketing smokescreen just before Comdex - where Broadcom says they'll be demonstrating the Atheros problem."

Comments

FCC?

[pvt_medic]

Doesnt the FCC have to approve such item to be for sale in the US market, And part of the approval process is to check to see if the decive interfears with other electronic devices?

Re:FCC?

[Anonymous Coward]

No, 2.4GHz is part of the unlicensed band.

Re:FCC?

[pvt_medic]

Having been bothered enough by my question I went and looked it up. I first looked at a walkie-talkie i had and Read the following "This Device complies with part 15 of the FCC rules. Operation is subject to the condition that this device does not cause harmful interference."

Then looking up Part 15 of FCC Rules available HERE. I focused in on 15.5 General conditions of operation.

Which states

TITLE 47--TELECOMMUNICATION

CHAPTER I--FEDERAL COMMUNICATIONS COMMISSION

PART 15--RADIO FREQUENCY DEVICES--Table of Contents

Subpart A--General

Sec. 15.5 General conditions of operation.

(a) Persons operating intentional or unintentional radiators shall not be deemed to have any vested or recognizable right to continued use of any given frequency by virtue of prior registration or certification of equipment, or, for power line carrier systems, on the basis of prior notification of use pursuant to Sec. 90.63(g) of this chapter.
(b) Operation of an intentional, unintentional, or incidental radiator is subject to the conditions that no harmful interference is caused and that interference must be accepted that may be caused by the operation of an authorized radio station, by another intentional or unintentional radiator, by industrial, cientific and medical (ISM) equipment, or by an incidental radiator.
(c) The operator of a radio frequency device shall be required to cease operating the device upon notification by a Commission representative that the device is causing harmful interference. Operation shall not resume until the condition causing the harmful interference has been corrected.
(d) Intentional radiators that produce Class B emissions (damped wave) are prohibited.

Re:FCC?

[afidel]

Not really, the 2.4Ghz spectrum that the .11b and .11g protocols use is an ISM band which is basically the FCC's term for junk spectrum that can be used for virtually anything. This is unliscensed space and the rules are pretty lax. So long as you aren't exceeding power requirements and aren't bleeding into adjacent frequency spaces you are pretty much ok. The frequency seperations used in .11b and .11g are IEEE standards not FCC ones.

That being said we observed this with Atheros's .11b channel ganging tech. Not only were they using two channels but their side interference in that mode went from bad in their normal mode to absolutely atrocious in the "turbo" mode.

Re:FCC?

[Spazmania]

That means that you don't have to be licensed to *use* equipment that operates in that band. The systems themselves still have to go through certification testing at a licensed lab to make sure they comply with the rules for a given band.

Re:FCC?

[squiggleslash]

This is true, but in response to the original question, the rules that apply have to do with frequency usage, duration of transmissions, and maximum power. They have nothing to do with the protocols you use.

Re:FCC?

[GigsVT]

Yeah but look at part (b). You also must accept interference. Two people with Part 15 devices that interfere with each other have to work it out between themselves. The FCC only gets involved if a Part 15 device is interfering with licensed users, like hamradio, police radar, TV, commercial radio, etc.

Re:FCC?

[Spazmania]

Yes and no.

Yes they do, but the FCC has specified that a device may use the whole spectrum up there. The 11 channels are an IEEE standard, not an FCC rule. As long as it uses spread spectrum, is under the ERP limit, and has been put through the appropriate certification testing, it can use as much or as little of that frequency band as it wants.

For example, FHSS radios (like 802.11a) DO use the entire spectrum. If you use an 802.11a radio near an 802.11b system, it'll stomp all over the 802.11b system severely degrading performance. Also, think: 2.4ghz range extender telephones. These don't use channels the way 802.11b does either.

Also note that even under the IEEE's channels, the side by side channels have sidebands which stomp on each other. That's why your manual tells you to leave an empty channel between the one you're using and the ones that other 802.11b networks in your area are using.

Re:FCC?

[AndroSyn]

Well you are wrong about 802.11a clobbering an 802.11b system. 802.11a is in the 5GHz range not 2.4Ghz.

Re:FCC?

[Spazmania]

Whoops, you're right, 802.11a isn't in the same frequency range as 802.11b.

The original point is valid though: frequency hopping spread spectrum radios (FHSS) will not coexist in the same spectrum as direct sequence spread spectrum (DSSS) radios. They clobber each other. And there are FHSS devices sold which use the 2.4ghz spectrum like 802.11b... Even if they're not 802.11a.

Super-G FaceOff!

[Tumbleweed]

Okay, so what happens when two Super-G networks are within interference distance of each other? Do they both drop down to 54Mbps speed, or what?

Re:Super-G FaceOff!

[ShawnDoc]

I can confirm the story. We recently tried testing out the G+ at work. We have two B APs, on channel 1 and 11. The G+ mode only works on channel "6". When we turned on the G+ network both of the B APs for all intents and purposes were unreachable. The G+ network functioned, but at slower than G speeds. The only way we could get the G+ to work was to turn off both B APs.

I'm going to assume that rather than using the real channel 6, G+ uses maybe 3 and 9. I say this since to make G+ work, it has to broadcast two different streams at the same time. Since channel "6" is the only option for G+ mode, I'm guessing they actually uses in between frequencies to avoid overlap between the two.

Guilty Party

[xwred1]

I'd be inclined to side with Atheros if only because I hate Broadcom. Atheros actually makes effort to get their chipsets supported on OSS operating systems, Broadcom does not.

Re:Guilty Party

[TWX]

Worse than that, Broadcom is likely responsible for using Linux and not releasing their modifications to the source code to those who have purchased the product (a base station) containing the GPLed code. So on top of Broadcom playing a marketing game, we have more reason to hate them.

I suspect that Broadcom's network interface module wasn't written right, and rather than be a complete non-depending piece of code, it depends on something that's been GPLed or is part of the kernel itself, and that would mean that it too would have to be GPL. If it is GPLed, then Broadcom would be releasing specifications on their devices that they don't want us to have, and we wouldn't have to pay for the base station anymore, just for the card...

Re:Guilty Party

[Tony Hoyle]

Broadcom *do* produce OSS drivers - just not for their 11g stuff. I'm told that's due to patents (not sure how true that is, though).

They also released the source to their DOS configuration utility to their network cards (the files had 'copyright broadcom' and 'not for redistribution' written on them, however).

Re:Guilty Party

[bfree]

I'd thought the reason for Broadcoms complete lack of a Free driver (you can use linuxants ndis driver loader with the windows drivers) for their wlan cards is that their cards are in fact general purpose radio transmitter/receivers, and with an open driver these features could be exposed and used which would prevent them selling the same hardware with different drivers for other applications. Of course this could be completely wrong!

Re:Guilty Party

[jhunsake]

Isn't it possible that what both of them are doing is wrong. Shouldn't we be getting on both their cases about different things?

You sound like child.

Re:Guilty Party

[Anonymous Coward]

1. Some of the displeasure with Broadcomm in the OSS community concerns the use of Linux in their base-stations without respecting the GPL license that accompanies it. If they did take GPL code (as some have claimed, I have no knowledge of this) to create a product and are making money off of it without respecting the license, that is wrong! and no amount of "wordly experience" excuses that!
2. Most of the OSS community has never asked manufacturers to create drivers for their chipsets for OSS; they merely ask that they release specifications for their chipsets to allow them to write drivers! Which you think would be a financial incentive for them.
3. It is certainly not naive to express your displeasure with a company's policies by "voting with your feet"; i.e. taking your business elsewhere.

Re:Guilty Party

[localghost]

Stop being ignorant.

Re:Guilty Party

[cymen]

Linksys And The GPL: The Saga Continues

Possible Linksys GPL Violations: The Saga Continues

more on kernel mailing list

*cough*

Re:Guilty Party

[cr0nj0b]

what do they contribute to OSS?

I have a linksys card that I've had for about 2 years. It has a broadcom chipset and they've never released a driver or plan on it. They use linux in their wireless Access Points and dont release the drivers they use there. I highly recommend never buying anything with a broadcom chipset. EVER!

The "not my problem" problem

[SeanTobin]

Don't you just love the joys of unregulated spectrum usage? If someone can get 108Mbps out of thier wireless by using more spectrum, I say let them do it. As long as its within power regulations, I can fine tune my wireless around it. If I get really upset I suppose I can buy one of the super-g systems for myself.

When the FCC sais that anyone can use this swath of the spectrum for anything within these power levels, and someone makes a gadget that does so, people have no right to complain if it interfears with thier gadget.

Re:The "not my problem" problem

[jerde]

Exactly!

This is just like arguing that two stereo systems in the same room "interfere" with each other.

Have these people forgotten that certain 2.4GHz cordless phones completely blast 802.11 devices off the air?

I, for one, would like to see a system that uses the entire spectrum to allow upwards of 1000Mbps throughput. There are certain places where there'd be an application for that use of the spectrum.

- Peter

Re:The "not my problem" problem

[Jameth]

To the contrary: People have every right to complain, but no right to legal action.

This is a very distinctive difference, because bad press and legal action are about equally harmful in the US.

Of course, the bad press might do nothing. However, it might do something.

Re:The "not my problem" problem

[TheLoneDanger]

This is a very distinctive difference, because bad press and legal action are about equally harmful in the US.

I think that bad press is more harmful than most legal actions. Legal action is pretty much guaranteed for a company of any real size and most have their own lawyers ready to deal with it. Stuff can be delayed for years until it doesn't really matter to anyone any more, and there's so much legal stuff all the time that most people ignore it.

However, bad press can create immediate and lasting negative impressions of a company (even if the company isn't actually at fault). I think it's rare that bad press does no damage (hence the bad part). Of course, some legal actions can create bad press...

Re:The "not my problem" problem

[Lord_Dweomer]

Before you just assume that an unregulated spectrum is a good thing, I would suggest you read up on The Tragedy of the Commons. It is an economic theory that most certainly applies to your post.

Sucks for them

[autopr0n]

IIRC, the only regulations for those frequencies are power restrictions. Beyond that, you can do whatever you want. Even if it breaks everything else.

It may be that the FCC would consider two seperate broadcasts can only use half the power each, but I kind of doubt it.

Two words...

[EmagGeek]

So what!? It's unlicensed spectrum and they can do with it what they wish as long as they don't violate any FCC rules, which Atheros is clearly not.

Part 15 devices must accept any interference, including that which may cause undesired operation.

I'd personally like to see point-to-point 2.4GHz hardware that uses the ENTIRE spectrum for extremely high bandwidth applications... that'd be very cool

Re:Two words...

[squiggleslash]

So what!? It's unlicensed spectrum and they can do with it what they wish as long as they don't violate any FCC rules, which Atheros is clearly not.

Wait, because they can you believe they should?

Part of living in a harmonious society involves cooperating with others to ensure what you're doing does not impair the freedom of others. This, at a very high level, is one of the reasons why we go around inventing protocols - clear, agreed-upon, protocols ensure that everyone can benefit from a technology rather than whoever has the most costly design.

There's a difference between what's legal, and what's right. Damaging the ability of others to use an established technology for minor bandwidth gains on your part is a Tech-world example of what may be legal but what most certainly isn't a decent thing to do.

Re:Two words...

[Wesley Felter]

Big corporate IT departments are too dumb for that; they demand to be locked in to one vendor (usually Cisco, because they sell the oldest and most expensive wireless equipment).

But you know you can turn turbo mode off, right?

Re:Two words...

[EmagGeek]

THey have every right to use the entire spectrum if they so desire. If you want to start talking about ethics and sharing, maybe you should stick to licensed spectrum. There are already provisions in the law regarding ISM bands that ensure equitable sharing. They're using 2 adjacent channels - which leaves 9 for others to use. Also, the power limit is the same regardless of bandwidth, so they have half the spectral power density than a single-channel transmitter, which means their range is far less.

Broadcom has a burr up its ass because they didn't think of it first. Boo hoo.

Re:Two words...

[BZ]

> They're using 2 adjacent channels - which leaves 9
> for others to use

Actually, no. They're using 5 and 6, which overlap with 2, 3, 7, 8 if the device is operating perfectly (the channel bands just overlap by design). If the device actually has out-of-band leakage as the article claims, the could easily be wiping out all 11 channels.

Re:Two words...

[squiggleslash]

Broadcom has a burr up its ass because they didn't think of it first. Boo hoo.

Why should Broadcom care if they thought of it first?

The issue here is that it appears that Atheros are producing networking equipment that makes existing 802.11b/g equipment useless - so your comments about what the law says they can do, and the comment about them "only" using two channels, is hardly relevent here. Building something you can easily see mucks up the existing standard is bad citizenship. Again, just because you have a right to do something, doesn't make it the right thing to do.

Why didn't Atheros just avoid that whole area? Why not use one of the other ISM bands?

Re:Two words...

[cyberformer]

They do. Atheros's main expertise is in 802.11a, which does use the other band and has more than 100 channels (with about 12 non-overlapping). They've been shipping a dual-chanel chip in that band for more than two years.

If I was a conspiracy theorist (and I am), I'd say that Atheros is deliberately breaking .11b and .11g equipment so that people will upgrade to .11a instead.

Re:Two words...

[davidstrauss]

I'd personally like to see point-to-point 2.4GHz hardware that uses the ENTIRE spectrum for extremely high bandwidth applications... that'd be very cool

Analogy: I'd personally like to see a widget the color red that is every color for extremely high bandwidth applications... that'd be very cool

2.4GHz is a specified part of the spectrum. Do you perhaps mean "every IEEE 802.11b/g channel"?

One could also interpret your statement to mean an UWB device that happens to include 2.4GHz in its range, but I doubt that you mean that because then the 2.4GHz means says little about the UWB hardware. Such an interpretation would be analogous to specifying that a full-color LCD displays red.

Not much

[autopr0n]

IIRC, there are at least 11 diffrent Wifi Channels to use.

Re:Not much

[CTho9305]

I think you can't use them all - don't they overlap? I think you can only use 3 or 4 at a time.

Re:Not much

[inburito]

Yeah and only three of them can be used simultaneously without any interference. Basically any given channel will overlap with the four neighbouring channels(two in both directions).

Different bandwidths?

[ActionPlant]

I don't understand how, on a different bandwidth, this can cause a problem. Even if they're using two different channels, the phasing doesn't make sense to me.

What am I missing here? It looks suspiciously like a last-ditch ploy to try to take out a competitor who offers a better product.

Rather like the throes of death, although it's a little early in the game to predict that.

Damon,

Re:Different bandwidths?

[eggboard]

If you look at the radiation pattern for Wi-Fi (or any spread spectrum) it has signal concentrated over a range, but there are harmonics and side lobes and out of band leakage -- it has to be within certain tolerances, but it's allowable. So if you're centered on channels 5 and 6, there will be slop into 1 and 11, but most devices are designed to have the smallest amount of slop.

Spectral Efficiency

[yogensha]

Alot of gear operating in the 2.4GHz area uses the _whole spectrum_. Western Mux (Proxim?) makes Wireless T1 gear which seems to be very popular for backhauling traffic from Cellular sites. Power output is regulated, but not spectral efficiency.

Social trap.

[SharpFang]

I live on a suburb with several (20+) radiomodem access points - no cables, no ground lines, just several "clusters of houses" hooked to their antennas. Connections to several ISPs.

The network connection sucks.

At first it worked fine. I'd say it worked great. People heard it works great so they began installing the equipment themselves. The lines began disturbing each other, but it still worked okay. More people installed this, and the network quality began to suck really. So some of them, to overcome the noise, installed signal amplifiers for their antennas. Result? Everyone without amplifiers simply lost their connections completely. So people began installing amplifiers en masse, which resulted in that connection sucks for everyone again. My packet loss ranges from 10 to 60%. TV signal gets disturbed. Radio mice and keyboars don't work. Great, just great. And the ISPs just can't come to agreement on putting one, good, shared ground line.

Ether is a limited resource and wireless in larger amounts will suck, no matter what.

Re:Social trap.

[Bishop]

You probably have a legitamate complaint for the FCC.

Re:Social trap.

[AmigaAvenger]

So people began installing amplifiers en masse,

get rid of your amp first, then complain to the FCC. They WILL do something about it, like fine anyone with an amp $100,000. Most people don't know that you are quite restricted as to what you do with 802.11. Adding a pringles can antenna in most cases is even illegal.

Re:Social trap.

[barc0001]

Easy fix: Get an old microwave, and build an EMF gun with it, then drive around and "stress test" some of your neighbors' antennas until your signal improves.
Rinse, repeat as needed.

Or, for something that won't get people upset, start a neighborhood committee and try to work out an effective way to turn this ad-hoc madness into something with structure yourselves. If it turns out you can get everyone going through 2 or 3 antennas instead and everyone is subconnected with wifi or even cable runs, and everyone's sharing the costs of the connections, everyone's connectivity will improve, and access costs will drop.

But the big trick is getting everyone on board. Though nothing is a motivator like "Hey! Want your Intar-web to work better and cost less at the same time?"

Re:Social trap.

[SharpFang]

Non-US. Other regulations, other laws. Plus by removing amps, any ISP brings all its customers offline completely. That means more or less death to them, at least in my area.

Re:Social trap.

[Rasta Prefect]

Adding a pringles can antenna in most cases is even illegal.

ICBW, but I believe the FCC restrictions are on total output power - How you choose to direct that is totally up to you, leaving pringles cans in the distinctly legal catagory.

In response to the parents post about about amplifiers, most amateur installed amplifiers tend to be crap. They tend to spew radio energy all over the spectrum. I've seen amps that actually output less at the desired frequency than the unamped equipment. This would certainly explain problems with otherwireless devices, such as TV's.

Re:Social trap.

[batura]

God man, you just summed up the cold war while talking about radio modems =)

Re:Social trap.

[jonbrewer]

I live on a suburb with several (20+) radiomodem access points - no cables, no ground lines, just several "clusters of houses" hooked to their antennas. Connections to several ISPs.

The network connection sucks.

Where is this mythical place with no landlines and twenty wifi connections?

Re:Social trap.

[imsabbel]

put the magnetron out of the microwave (thats the thing located near the opening in the inside). Use a sattelite dish of at least 75 cm ( 1m is better), put the magneton in the place of the reviever, opening facing the dish.

And then just power up and aim.

have fun... (a 1000W magneton should be able to kill a high gain antenna of an acess point from at least 50m distance....

Re:Social trap.

[imsabbel]

no, its not. Its a limit on what total power equivalent is distributed in any direction.

Re:Social trap.

[mlyle]

No, it's based on EIRP. Effective isotropic radiated power. In other words, the amount of power sent in a given direction as would be radiated in that direction by an ideal radiator with a given wattage.

100mW + 10dB of antenna "gain" = 1W EIRP.

Of course, pringle can antennas don't provide much gain, because they're too small to be effective waveguides for 2.4GHz. Most of their benefit is in noise rejection. I have not busted out the spectral analyzer myself on this topic, but I doubt they'd bump a normal wireless card over the legal limit of EIRP.

Re:Social trap.

[way2trivial]

I had this happend to a gentleman whose computer I support.. we went powerline..
worked great...

Re:Social trap.

[Salamander]

Welcome to Garret Hardin's "tragedy of the commons". People occasionally cooperate voluntarily to manage a shared resource, but much more often each person just grabs whatever they can and nothing short of physical force or some other means backed up with the threat of physical force (e.g. a lawsuit) will make them do otherwise. You have just illustrated how laissez-faire fails in the real world.

Just a quick question

[Naffer]

What limits throughput on a wireless connection? Is it broadcast power, frequency, amount of spectrum, or "c"?

Re:Just a quick question

[SharpFang]

Background Noise.

The higher the frequency, the better "available bandwidth" (signal wave must be relatively low frequency to carrier wave - you modulate the carrier wave), but then the stronger the noise to signal ratio. You can apply different counter-measures against it, but all they have a very limited effect and sooner or later it boils down to "increase signal power". And once the signal power gets increased, it adds up to other signals noise. Which forces them to increase their power. Which adds to total noise level forcing yet more to increase the power... Deadlock.

Re:Just a quick question

[mlyle]

All of the above!

First, there's fundamental thermodynamic/information theoretical limits on the amount of information that can be carried in a given amount of spectrum for a given number of symbol values. This is because any "modulation" of a carrier causes the carrier to have spectral products off the carrier frequency. Modulating the carrier at faster rates makes the overall signal wider. I could go deep into sampling theory here but I won't.

The wider a signal in spectrum, the more noise there is over that spectrum, and the more power must be spread over that spectrum to be detectable over the noise floor.

Maximum Theoretical Throughput in BPS = Bandwidth * Log2 (S/N+1)

Where S is the signal strength and N is the total amount of noise in the signal bandwidth. Note that N increases linearly with bandwidth in typical RF applications.

Frequency also matters because the amount of noise present on different parts of the spectrum varies. Almost all noise above 100MHz, though, is created either thermally inside the receiver.. or comes from man-made sources.

Re:Just a quick question

[mlyle]

Oh, yah, I forgot to discuss C.

All the consumer 802.11* products today are half duplex. That is, they can't transmit at the same time they're receiving, and only one party can talk at a time while still allowing the signal to be demodulated successfully. (Vivato is doing some really neat work with phased arrays to receive from multiple people at once, but that is high-end $10K+ hardware).

Channel arbitration, or deciding who can talk when, is expensive, and isn't perfect. A certain amount of the time, two wireless transmitters will decide to talk at the same time.

802.11 uses a CSMA/CA protocol to prevent this "doubling". CS is Carrier Sense-- side A can tell when the other side is transmitting.. but in this case usually only when side A is not transmitting itself. MA indicates that it is a multiple access channel. CA=collision avoidance specifies that a protocol is involved where one side transmits for a little while, and then stops and checks to make sure there's no one else transmitting.

The amount of time needed to spend on the test carrier is related to the propagation delay between stations-- in other words the speed of light and distance, as well as the switching times from transmit/receive and the amount of listening time necessary to determine if someone else is transmitting, as well as various other probablistic answers (the odds of someone else both starting to transmit and also checking for someone else at the same time, for instance-- adjusted for propagation delay, of course).

All of this arbitration, waiting, etc, wastes bandwidth, and thus lowers real world throughput.

Re:Just a quick question

[jerde]

What limits throughput on a wireless connection? Is it broadcast power, frequency, amount of spectrum, or "c"?

All of the above!

The most limiting factor is the amount of spectrum -- the bandwidth allocated to the signal. If you're restricted to a very narrow set of frequencies, you can't send as much data in a given amount of time.

But, once you've picked the bandwidth to use, broadcast power and frequency each affect how well the transmitted signal can be received. Lower frequencies pass through solid objects more easily, for example, but often require more broadcast power. There are myriad details affecting the quality of a signal as it passes from the transmitter to the receiver, not the least of which is interference from other transmitters.

And finally, "c" ultimately affects the latency of your signals. NASA's conversations with the Voyager spacecraft would be an extreme example: it takes about 25 hours now for a signal to make a round trip from Earth to Voyager 1 and back. Closer to home, even Earth satellites are far enough away that the speed of light becomes a limitation -- the round trip to a geosynchronous satellite is 0.239 seconds. If your signal is mostly one-way, this has very little impact. Television, for example, isn't hindered by this. But any kind of interactive connection like online gaming or even telephone conversations are adversely affected by that amount of latency.

I still think that the emerging spread spectrum technologies are our best bet for wireless computer networking.

- Peter

Re:Just a quick question

[dougmc]

I still think that the emerging spread spectrum technologies are our best bet for wireless computer networking.

I agree. But emerging? They're already here!

What do you think 802.11(whatever) and WiFi is? :)

How about D-Link?

[danielsfca2]

I don't know about Atheros, but ever since my 2-floors-down-neighbor put up a D-link "b" access point, I can't connect to my bedroom (g) access point from my own living room, even though I've separated the channels (2 and 11). I am forced to connect to his.

Does anybody test for interference with these damn things?

Re:How about D-Link?

[danielsfca2]

> a scrambler so that no one in the condo can use the spectrum?

Apparently you can. It's called the D-Link Wireless-B Access Point.

I use cables for lots of things, but it really takes the utility out of my AirPort card that I can't ever use it in the living room. I used to use it in there all the time; now I'm stuck at my desk.

Re:How about D-Link?

[FroBugg]

If you don't mind the cancer risk, you could just rip all the shielding out of a microwave and keep it running. That should do a pretty good job of it.

Re:How about D-Link?

[aminorex]

Pull up the carpet, apply aluminum foil,
replace carpet.

I wonder if...

[Dead_Smiley]

this is why my Dlink G network stopped working one day a couple of weeks ago. I was working just dandy for over two months and then it would only connect at 1-2 Mb. The connection would only last a minute or two and then nadda.

Atheros Linux driver

[bersl2]

here.

For the sake of relevancy.

Also, almost everything is GPL'ed. So: Atheros == good && Broadcom == bad;

Re:Atheros Linux driver

[kju]

Your post ist not Insighful, but Clueless. So you say almost everything is GPL'ed. Yes, exactly, with the totally small exception of the COMPLETE hardware abstraction layer. Yes, everything which deals with the hardware is binary only and NOT gpl'd. The rest of the driver is worthless, the heart and core would be the HAL. So this driver is better than nothing but not really acceptable.

Atheros is as bad as Broadcom is. Dump both and buy Intersil PrismGT. Not even did they support a linux driver effort, they even sponsored the developent!

Re:Atheros Linux driver

[The Vulture]

Also, almost everything is GPL'ed. So: Atheros == good && Broadcom == bad;

Normally I'm inclined to agree, but not in this case. Having worked for companies that have dealt with Broadcom, and talking to some of their engineers myself, I know a bit about the company itself.

Their engineers are very smart people (most of them hold degrees above the typical Bachelor of Science), and I've seen their source code (for cable modems), and it's well written. Their spec sheets (again, for their cable modem products) are generally top-notch, although I'm a bit miffed at them about the whole Broadcom 3415 tuner chip issue (with the patent infringement suit against Microtune), and the whole Linksys/Cisco situation. I'm also not keen on their interview processes either (myself and a couple of my former co-workers have all interviewed there in the past), but that's not related to this issue.

That aside, I'll address the statement from you that I copy/pasted... If the hardware sucks, having GPL'd drivers for it is of no use (at least not to me). If the Atheros chipset is causing interference (and I reserve my judgement until I hear more about the issue), then as far as I'm concerned, I'd rather buy other hardware that actually works.

-- Joe

Re:Atheros Linux driver

[Jameth]

Nice limited view of reality there

OSS == Good && Proprietary == Bad ?

How about this:

Crap == Bad && Proprietary == Bad && OSS == Good && Reliable == Good && Polluting == Bad && ellipsis

Try to look at good/evil/right/wrong/good/bad/positive/negative in a slightly broad view. There is more to the world than OSS and the GPL.

Re:Atheros Linux driver

[localghost]

Just because it's got GPL drivers doesn't mean it's good. Here's an excerpt from my kernel log:

Nov 15 15:39:08 debian kernel: ath0: hardware error; resetting
Nov 15 15:40:00 debian kernel: ath0: hardware error; resetting
Nov 15 15:47:15 debian kernel: ath0: hardware error; resetting
Nov 15 15:48:18 debian kernel: ath0: hardware error; resetting
Nov 15 15:49:07 debian kernel: ath0: hardware error; resetting
Nov 15 15:56:21 debian kernel: ath0: hardware error; resetting
Nov 15 16:02:18 debian kernel: ath0: hardware error; resetting
Nov 15 16:04:03 debian kernel: ath0: hardware error; resetting

Of course, that beats my Prism 2.5 802.11b card:

Oct 21 17:05:38 debian kernel: eth0: Error -110 writing Tx descriptor to BAP
Oct 21 17:06:09 debian last message repeated 38930 times
Oct 21 17:07:10 debian last message repeated 77347 times
Oct 21 17:08:11 debian last message repeated 76456 times
Oct 21 17:09:12 debian last message repeated 76904 times
Oct 21 17:10:13 debian last message repeated 77600 times
Oct 21 17:11:13 debian last message repeated 75482 times
Oct 21 17:12:13 debian last message repeated 72258 times
Oct 21 17:13:13 debian last message repeated 70948 times
Oct 21 17:13:47 debian last message repeated 39696 times

The Prism 2.5 card has since behaved better, but the Atheros one is leaving those messages constantly.

Re:Atheros Linux driver

[The Salamander]

Is it really an FCC issue as their documentation explains? does this
mean you can't get specs for software radios?

This code manages much of the chip-specific operation of the Atheros driver.
The HAL is provided in a binary-only form in order to comply with FCC
regulations. In particular, a radio transmitter can only be operated at
power levels and on frequency channels for which it is approved. The FCC
requires that a software-defined radio cannot be configured by a user
to operate outside the approved power levels and frequency channels.
This makes it difficult to open-source code that enforces limits on
the power levels, frequency channels and other parameters of the radio
transmitter. See

http://ftp.fcc.gov/Bureaus/Engineering_Tech nology/Orders/2001/fcc01264.pdf

for the specific FCC regulation. Because the module is provided in a
binary-only form it is marked "Proprietary"; this means when you load
it you will see messages that your system is now "tainted".

If you wish to use this driver on a platform for which an ath_hal
module is not already provided please contact the author. Note that
this is only necessary for new _architectures_; the HAL is not tied to
any specific version of your operating system.

Re:Atheros Linux driver

[The Vulture]

End-users probably won't get the specifications. However, if the company I work for (for example) signed the pre-requisite paperwork and purchased significant product, we'd not only get the spec sheets, we'd probably also get the source.

Of course, neither of those can be leaked outside of the company without civil reprecussions.

I believe the problem stems from the fact that the hardware manufacturers would rather leave the hardware open-ended and cap it's capabilities in the software - this makes it much easier to issue upgrades (where necessary), and leverage parts for future products. Plus, if the FCC rules change, it's much easier for the manufacturer to comply.

As a good example, a product from the previous company I worked for, Com21. The DP1010 cable modem (which was pretty popular about five years ago, or so), rather than implementing packet filtering logic in software, implemented it in a custom ASIC. When CableLabs, back in the time of Certification Wave 12 (I think it was CW12) implemented the DOCSIS Cable MIB (RFC2669), the DP1010 could not provide this functionality (because the actual hardware did not allow it). Many customers (and even Com21 to some extent) got screwed over on this, as RFC2669 is what provides the method of most of the security of cable operators nowadays.

-- Joe

Re:Atheros Linux driver

[fishbowl]

>Buy Intersil PrismGT

That's easily said. Exactly which cards are those? We need to know BEFORE we buy the card,
and we need to be assured of it.

Re:Atheros Linux driver

[fishbowl]

>As for 11g, all of them I've seen are Broadcoms.

Most of the replies to my queries have been from people who don't realize how bad the situation really is. People who already have wireless cards that work, don't understand that it is impossible to go to a retailer today, and purchase a card that is known to be compatible. I was one of them. Linksys was Prism2, the usb device was wonderful, so what's the problem? Well, my 3rd one wasn't Prism2, despite having the same packaging and model number. So much for my relationship with Linksys.

There might be 802.11g cards that work with linux, but I don't see them in any retail store I know about, and I don't see any way to know for sure, BEFORE buying one.

Re:Atheros Linux driver

[Ewan]

http://prism54.org/

There aren't a hell of a lot of them, but that site tells you which ones they are.

Ewan

just move to the 5 GHz band

[partha]

With 12 more channels as of this past week, why would one fight over the crowded 2.4 GHz band? Unless one doesn't have a product for the less crowded 5 GHz band.

Re:just move to the 5 GHz band

[Geekbot]

"why would one fight over the crowded 2.4 GHz band?"

Because I want equipment that will be able to penetrate through a couple of walls and a floor or 2.

This is the very attitude...

[SharpFang]

which leads to killing the wireless transmission.

See my post a bit below, and try to understand - if you don't give a damn, you will just force people to move to equipment like you have. And would you like to have 1K/s transfers like I have? Or days with ping like this?

64 bytes from [provider]: icmp_seq=801 ttl=61 time=179.688 msec
64 bytes from [provider]: icmp_seq=1372 ttl=61 time=69.358 msec
64 bytes from [provider]: icmp_seq=2020 ttl=61 time=139.666 msec
64 bytes from [provider]: icmp_seq=2411 ttl=61 time=29.708 msec
64 bytes from [provider]: icmp_seq=2735 ttl=61 time=29.201 msec
64 bytes from [provider]: icmp_seq=2808 ttl=61 time=119.482 msec
64 bytes from [provider]: icmp_seq=3540 ttl=61 time=1.239 sec
64 bytes from [provider]: icmp_seq=3869 ttl=61 time=29.728 msec
64 bytes from [provider]: icmp_seq=4420 ttl=61 time=59.605 msec
64 bytes from [provider]: icmp_seq=4814 ttl=61 time=29.677 msec
64 bytes from [provider]: icmp_seq=4877 ttl=61 time=129.729 msec
64 bytes from [provider]: icmp_seq=4974 ttl=61 time=59.681 msec
64 bytes from [provider]: icmp_seq=5243 ttl=61 time=269.733 msec
64 bytes from [provider]: icmp_seq=5645 ttl=61 time=19.729 msec
64 bytes from [provider]: icmp_seq=6015 ttl=61 time=399.669 msec
64 bytes from [provider]: icmp_seq=6263 ttl=61 time=69.728 msec
64 bytes from [provider]: icmp_seq=7126 ttl=61 time=49.671 msec
64 bytes from [provider]: icmp_seq=8205 ttl=61 time=139.727 msec

Re:This is the very attitude...

[Rev.LoveJoy]

I am in a very similar situation to you (I have read your other posts under this article). I live in a rural area with a large number of wireless providers. Early on, access was fantastic - then mom & pop internet shops figured that by spending $300 a month for tower space and a bit more on equipment they could become "Wireless internet providers."

The story goes just like yours; about half the time my transfer rates are 1Kbps and below. Making matters worse are my neighbors whose shoddy installs leave their PCs polluting what little wireless bandwidth is available with ARP and NETBIOS garbage packets (actually, a signifigant percentage of the total traffic over a given 24-hour period will be arp spam and browser announcements).

The tactic I have taken is to petition the mom & pop providers to clean up their act via a professional organization to which I belong (for network geeks). This has been met with limited success. I have not been at it for long. If this tactic fails, I may go with the other poster's advice about the old microwave.

Cheers,
-- RLJ

omniscient article

[zymano]

Will Success Spoil Wi-Fi?

Re:omniscient article

[joshtimmons]

I would like to read an omniscient article. Alas, that one is only prescient.

It's not unregulated.

[mindstrm]

It's unlicensed, there is a big difference.

You cannot just use this spectrum however you see fit, there are rules, quite serious ones.

You have every right to complain if the device in question is illegal according to the regs.

How about this..

[little alfalfa]

How about Broadcom release the specs of their chips so that we can have more Linux Wi-fi support. It kills me that I have a card with one of their chips on it and it is USELESS to me.

Re:How about this..

[r_cerq]

It isn't useless, just not open.
It's been discussed a few times, but I'll say it again: you CAN use a bcom chip in Linux, if you buy Linuxant's driverloader. My latest laptop has a bcm4301 builtin; I stopped carrying the Prism2 PCMCIA around once Linuxant's driver appeared: I ended up buying a license for it last week, since it works fine and saves me the trouble of carrying yet-another-PCMCIA.

Everybody can

[Levvie]

afaik, the only thing what happens using Atheros is that is bundles 2 channels. If I would buy two accesspoints, place them on a different channel, stick 2 cards into my notebook and assign two well balanced subnets to my home network, speaking in terms of air waves I did exactly what Atheros did, just on a lower level. This way I can transfer (multiple) data streams at double data rate without harming any rules what-so-ever, exept for my neighbours who have less channels to choose out ;)

Re:GWB's second term - don't let it happen

[DigiShaman]

As a republican, I can honestly say that GWB speaks from his heart. His decisions however, are not the wisest. The Patriot Act may not have been intended to take away our civilian rights in a malicious manor, but the foundation for it is there. And that's what truely scares me about the PA. It's possible that a future administration (regardless of party) could take advantage of this and totally start hacking away.

Improper channel useage is rampant in Wi-Fi

[craenor]

The 2.4GHz unlicensed band has 3 non-overlapping channels (1, 6 and 11). You can use up to three DSSS (Direct Sequence Spread Spectrum) devices in the same location without them interfering with one another to a great extent. This would include one or more Wi-Fi networks, 2.4GHz cordless phones (that use DSSS, not FHSS), Baby Monitor, etc.

Anytime you have more than three devices co-located some of them are going to interfere with one another. That interference is going to either degrade your connection speed or it's going to prevent you from being able to connect all together.

If this 108Mbps technology is truly setup to use channel 5 and 6, then Broadcom is right. It is going to interfere with 2/3's of the available non-overlapping channels.

On a quick side note, because wireless connections do not have collision detection, they have to rely on collision avoidance. Once a packet is sent the receiving station has to reply with a receipt acknowledgement before more data is sent, this basically works to cut the actual data transfer rate in half, not that it matters anyway, since almost all wireless networks are used for internet access from ISP's that are lucky to break 3Mbps.

Back to my point though...if you have a wireless network, then be a conscientious wireless user and keep to channels 1, 6 or 11. You can also use the site survey software that came with your wireless adapters to find out what channel other nearby users have occupied already so you can avoid those. Additionally, if you buy other wireless products, avoid the 2.4GHz band if you can. If you must get a 2.4GHz cordless phone or baby monitor then do your homework and strictly avoid those devices which use FHSS (Frequency Hopping Spread Spectrum) - they are frequency hogs which have a tendency to kill other wireless devices.

Craenor - Senior Wireless Networking Specialist for Dell, Inc.

Re:Improper channel useage is rampant in Wi-Fi

[FrostedWheat]

be a conscientious wireless user and keep to channels 1, 6 or 11

And don't use amplifiers! They cause more trouble than they solve. If you need a longer distance link, try larger antennas first.

Re:Improper channel useage is rampant in Wi-Fi

[WolfWithoutAClause]

Back to my point though...if you have a wireless network, then be a conscientious wireless user and keep to channels 1, 6 or 11.

It's not as simple as that though. The reason you have 11 channels is that in certain circumstances it's ok to overlap channels. For example, in an office block you might use 1,6,11 on the ground floor, and say, 4,9 on the second floor. Provided the flooring gives some degree of shielding, but not total shielding, the interference would be less than using 1,6,11 on each floor. The optimum pattern depends on the distances and materials used in the building; and that can only be determined by a site survey.

(For example, the last building I worked in had metal floors and metallic silvering on the windows, clearly 1,6,11 was fine on every floor.)

This is not new news

[puzzled]

If you place Adtran Tracers anywhere near an 802.11b cell the effect is rather like sandblasting a soup cracker. The Tracers split the band with one end using the lower half and the other using the upper. They bridge ESF T1 frames so their utilization of spectrum is always 100% whether they're idle or not.

Broadcom is just producing either a concatenated 108 mbits by using two channels at once or they're producing a full duplex 54 mbit 802.11g connection.

The FCC will not do anything about this sort of thing. As an unlicensed band user they'd prefer that you just drop dead.

Unlicensed band may work well in unpolluted rural areas with one carrier but in metro areas it is pretty much a disaster in the ISM band and the same troubles are starting to happen in the UNI (5.8) band as well.

If your business plan depends on flawless throughput in the ISM band please send me your home address - I'll come over, kick your ass in your driveway, steal a bunch of stuff from your house, and we'll call it good - the financial effects and suffering are the same but you get it all compressed into a few short minutes of fun, instead of spending a year of your life flushing your money along with investor's dollars.

Obvious solution

[Master_Wu]

Since probably at least half of your twenty neighbors plug the things in and go, just use their wi-fi points and vpn into your network if you need to - no bandwidth loss for you (it's all bonus, and free) and you can switch to the next guys nextwork if your current one gets slow.

Everyone...

[FrostedWheat]

Everyone say 'Awwwwwwwwwwwwwww' :)

Illegal modifications

[bobthemuse]

Some people have told me that it is possible to modify the hardware or purchase hardware from outside the US which uses different frequencies to get around this. Yes, it's illegal, but with this much RF pollution, would to be possible for the FCC to triangulate your location (say, inside an apartment building), and would they even care (assuming you weren't interferring with anything local).

OTOH, that happens when everyone starts doing this? God forbid when software radios become popular and people start trashing the entire usable spectrum?

Re:Illegal modifications

[Detritus]

Yes, it's illegal, but with this much RF pollution, would to be possible for the FCC to triangulate your location (say, inside an apartment building), and would they even care (assuming you weren't interferring with anything local).

They can do it. Whether they do it, and how fast they do it, has a lot to do with who filed the complaint and what type of interference you are causing. Fuck with the FAA's communication and navigation systems and you will be very sorry, very quickly. Plus, having flagrantly ignored the FCC rules, don't expect any leniency.

Re:To HELL with BROADCOM.

[craenor]

Dell still sells Truemobile 1150 mini-pci cards (which are only certified to work in Dell portable systems with a mini-pci slot) and Truemobile 1150 pcmcia cards. You can buy those through the Dell website or through sales.

They are both using the older Lucent chipset which has an abundance of available cross platform drivers (like Linux). On a side note, it is also probably one of the best wireless chipsets ever made, even if it is pushing 3 years old. These are 802.11b.

As for Broadcom not coughing up info or cooperating for Linux drivers...I've heard one explanation that makes sense. Their chipset is the same one used in all of their wireless products which also have military and industrial applications, and any opensource driver would allow users to adjust the chipset into frequencies not allowed by the FCC, for which Broadcom could be held responsible.

Is Comdex the right place?

[Slur]

Now as you see, when we flip this switch....

"Ah, fuck, my pacemaker!"

Re:GWB's second term - don't let it happen

[geekoid]

" As a republican, I can honestly say that GWB speaks from his heart. "

I didn't know republican could read each others minds.

To automatically say someone has good intentions just because the happen to be of the same party, is a mindless drone.

OK, I lose this one.

[bersl2]

Yeah.

I was kind of in a rush to post this one. Maybe I shouldn't have tacked on that last part... :-0

Gigabit

[AchmedHabib]

At least we got drivers for the gigabit ethernet ports. A good thing since now that they can be found in IBM's Intel servers. ;)

shouldn't this be modded +1 funny?

[ToKsUri]

because it is a joke, isn't it?

Why FrostedWheat is right about antennas

[KC7YRN]

Stock antennas are pretty unimpressive and leave a lot of room for improvement.

Antennas are cheap, especially if you build your own, and they don't burn up battery power.

Antennas work in two directions. An antenna with a better pattern improves your range for both transmit and receive. An amplifier on one side of a link doesn't help you hear the other side any better.

Antennas with radiation patterns that match where you need the network reduce interference coming in as well as interference going out.

Maybe mesh networks would help

[bonds]

Right now all of our wireless devices work on the "shout as loud as you can so people can hear you principle." When we get too many devices trying to work in the same space, it's like putting a bunch of people on opposite sides of the room and asking them to shout across the room to speak to each other. This works fine when one or two conversations are going, but it doesn't scale very well.

Mesh networks offer the possability of having each node pass a note to the node closest to them in the direction of the node they are trying to reach. They only have to speak loud enough for that closest node to hear, making meshes a lot more scalable. Like passing notes in high school rather than shouting across the room and getting the teacher pissed off at you. :) I suspect that as wireless devices become more popular we'll need something like mesh networks to make more efficient use of the spectrum. In fact, in a manner similar to Bittorrent and Freenet, the more people that participate in a mesh network, the more resiliant and speedy the whole network is.

sb

Re:Maybe mesh networks would help

[burns210]

sounds like IIP, Freenet, or just a wireless p2p protocol in general. This is a good idea, and one i have had for a while now. i wonder if porting freenet to a palm would work? and if coding the software, could the palmos start chatting to other palms without a problem? or are they hardcoded, somehow, to need an WAP?

any ideas?

Yet another perl hacker!

[gfilion]

Atheros' CEO is a perl hacker, so I guess that this would mean that whatever they do, it will be good, no?

Re:To HELL with BROADCOM.

[TO11MTM]

.... The Lucent Drivers SUCK for windows 2000. I kid you not, I had to talk to an orinoco rep because my laptop wasn't working with my Orinoco mini-pci card woth a crap. I told him I was running Win2k (Which is one of the OS options from the manufacturer...) And he immediately gave me the solution... Because it was a known problem. 1: Install Windows 2000 with NO service packs (This was when Win2k was being bundled with SP3, so I had to go hunt down a CD :)) 2: Install the M$ 802.11b Hotfix. 3: Install The Lucent Drivers. (You may have to install the Hotfix again) 4: Install Service pack 3. To quote him. "And that might work." Bloody Beautiful.

Re:To HELL with BROADCOM.

[fishbowl]

"Any opensource driver would allow users to adjust the chipset into frequencies not allowed by the FCC, for which Broadcom could be held responsible."

So, when someone in the Netherlands or New Guinea writes a driver, can they still be held responsible? If not, then the argument holds no water. If so, well, they should have thought about that before releasing the dangerous product to consumers.

Atheros vs. M$

[lwsimon]

Ok, so let me get this straight - Atheros is producing a WiFi AP that intereferes with all nearby APs, rendering them nearly inoperable... Microsoft ships Windows with IE, and makes it difficult for Netscape and similar to work properly... Whats the difference?

Re:GWB's second term - don't let it happen

[DigiShaman]

Yes, they will. After all, there is a war on drugs. Jesus, your such a fucking morron! *Smacks the back of your head*