Slashdot Mirror
Wi-Fi Cards Can Now Detect Microwave Ovens
An anonymous reader writes "Researchers at UW Madison have used regular WiFi cards to detect non-WiFi interference sources like microwave ovens, Bluetooth devices, cordless phones, Xbox controllers and video cameras. They call their software Airshark. Current products like Wispy, Spectrum Expert are expensive and need extra hardware, whereas Airshark is a software-only solution that can directly work on the Wi-Fi cards on your laptops and APs. This also paves way several interesting applications. For example, your WiFi network will not be affected anymore just because your neighbor switched on a microwave oven or a cordless phone — the newer WiFi APs will be able to switch the channels and adapt to the interference accordingly."
If you live in a city or in the suburbs, you can see LOTS of WiFi access points already, so switching to a different frequency won't get you to an interference free channel! Maybe you'll get to a little less populated one, but not interference free!
Then the entire spectrum problem is solved, and everything would be autoconfigured for the basic paradigm: connectivity. Now I don't expect a microwave to give me food-over-ip, but I would expect a neighbor wifi cell, helping my AP to extend the signal, if my client would move out of range (aka: has more noise).
Support Eachother, Copy Dutch Property!
"Honey, did you leave the microwave on?"
"I don't know, Dear, let me log into my PC and check."
The technology allows for this already. However, the security and privacy implications are big. Not to mention bandwidth limitations. And switching capability. And routing tables. And ARP tables. And those are the problems I though about while typing this. I'm sure there are several others.
morcego
RTFA. See a paper, but no link to binaries. Oh, "So far the researchers haven't commercialized Airshark.", right...
I thought that shielding was well understood and in fact a good reason of the part why microwave ovens are a common household item.
Could anybody with experience in these matters explain where the leak is coming from, and why do they still exist? Is it impractical or physically impossible to have perfect shielding for some reason?
When my downloads get slow and I can't refresh slashdot, it means it's time to take a break because mom is making me a snack upstairs.
Sig: I stole this sig.
"Hey I'm trying to stream a movie here! Oh wait it's the microwave? Can you bring me some popcorn?" 'Nuff said. ;)
Most microwaves oven are supposed to work right at 2450 MHz, so if you want to avoid microwave ovens (which you should if there's one in the area and it's used often), you want to use the highest or the lowest numbered WiFi channel -- you don't need a spectrum analyzer to tell you that.
But if you want to see the results of one on many microwave ovens ... here you go. It looks like the exact bandwidth used by their signals vary quite a bit, though my advice above still stands in the majority of cases.
Of course, there are other 2.4 GHz band users as well, and a scanner could be useful for pinpointing those.
(not to be confused with the WIFE Detector(tm) )
then you can indeed detect a microwave oven, and pretty much anything that spews out parasitic signals from 1500-3000 Mhz.
*Technical explanation coming up* ;)
This is due to the cheap construction of those So Called WiFi detectors, they're not digital, they're in fact analog receivers that only detect any modulation on the band (very VERY wide-band / BroadBand reception)... it's just a glorified Crystal Radio with a small half-coil, 3-4 transistors to amplify anything...any signal picked up by the small 1 cm internal antenna, and 1 transistor to switch on a led (or 3-4 resistors, if it's sophisticated and have 3-4 leds...ha ha)
There...now the Chinese can mass-produce them, I just literally gave you the schematics for it... ...oh wait!
What this world is coming to - is for you and me to decide.
I was going to say how this sounds like a potential invasion of privacy, but then I realized that the police knowing you have a microwave or Xbox controller probably isn't something to worry about.
The only escape is the 5 Ghz band, but if you could use that, you'd be doing so in the first place.
It's call mesh networking. If everyone did it, we wouldn't need telcos (or ISPs) so much.
So where's the source code for this?
Well, it's not just that, it's that the devices are designed to give the most range possible without going with a unidirectional antenna. And the problem is that it worked fine when the 802.11b devices were first rolling out because few people had them, but as they've gotten to be common, you then have to deal with a dozen WAP competing for scarce spectrum.
And all is well and good if you have a large property, but if you're in an apartment and just need something that's fast and can let you move from the desk to the kitchen table, say 15' away, having a device cranking out enough power to go 200' is way too much power.
802.11a devices (operating at 5.45 GHz) are already supposed to detect radar signals and switch channels if one is found. This is particularly a problem in Europe, where most weather monitoring radars are C-band, and share the same frequency band as 802.11a.
Not an expert on this, but couldn't the other hub just relay the encrypted signal, without dectrypting it? Then there is no security problem.
I would love to use my Bionic, headset, GV and sipdroid over wifi, currently not possible b/c of wifi interference.
sysadmins and parents of newborns get the same amount of sleep.
Bluetooth hops across the 2.4 Ghz spectrum. You can configure a Bluetooth host to avoid a given wifi channel, but it doesn't work the other way around. Again, the only way to avoid it is to not use the 2.4 Ghz band.
Custom kitchen, deliverieieies.
I had a microwave oven that consistently stopped my Netflix videos streaming over WiFi every time someone made a cup of tea.
I was able to prove a contributing issue was related to its poor door seal letting microwaves out using the free WiFi tool NetStumbler (Also known as "Network Stumbler").
NetStumbler has can graph the Signal/Noise ratio of a WiFi station over time. If you put a laptop running NetStumber in a microwave (Don't turn on the microwave!) you should see the signal to noise ratio drop 30 dBm as the door shielding attenuates the WiFi signal. If not, you probably have an old oven that has developed a wonky door seal.
In my case, I was able to feel the microwave door close a little more as I pressed the handle. And after alternating pressing and releasing the door without changing my body position, 10 seconds on 10 seconds off, I was able to clearly see a 5 dBm difference in the WiFi signal to noise ratio on my old oven. That didn't happen on my new oven.
I also saw other people comment that if a cell phone rings inside a microwave, then that's a sign the microwave is leaky. I doubt that's reliable, since many cell phones use a different frequency than microwave ovens. And they don't report signal strength accurately.
But can the researchers detect the government mind control rays transmitted from cell phone towers? If it could do that, I wouldn't have to wear this tin-foil hat all the time!
dragonhawk@iname.microsoft.com
I do not like Microsoft. Remove them from my email address.
Just stop using your microwave. There are few good reasons to do so.
Seriously we've become so fucking lazy that we can't even surf porn without being interrupted by a Lean Cuisine?
Good luck connecting across the Pacific or getting reasonable response times from half way across the continent then.
You are talking about point to point encryption. Yes, it would be possible, assuming you are able to establish encrypted point to point connections to all services you connect to. Or you can setup a VPN to a "relay" server. The first is simply not a reality at this point, and the second is something most people wouldn't know how to do.
Privacy would still be a problem, because anything up to Layer 3 would not be encrypted in the scenario you propose. Depending on how it is implemented, even Layer 4 would not be encrypted. Regardless, a lot of identifiable information (privacy issue, not necessarily security) would still be available.
morcego
Handy-dandy microwave signal detector! Hey, if your app can tell between wifi and microwave then that would be a neat thing to be able to track.
Hi, I'm working for The Serval Project, and like other projects related to wifi mesh routing, we do have high level goals like this. And we're actively trying to make them a reality.
One of our staff just returned from a presentation to IEEE, to propose a more open standard for the next 802.11 spec.
The basic premise of our proposal is that the protocol for using wifi devices to route traffic should be dealt with in kernel or user space. Not in the radio spec. And that adhoc, and 802.11s are useless for this task (Damn you BSSID, why you change?). We also think that security and perhaps even error correction should be dealt with via a VPN or baked into the application layer.
We want the next wireless spec to include a basic packet radio mode, operating in any unlicensed white-space spectrum, that gives as much control as possible to higher levels of the OS. So that new interesting ideas are easier to experiment with and implement.
And we've been invited to the next IEEE working group to help make it happen.
09F91102 no, 455FE104 nope, F190A1E8 uh-uh, 7A5F8A09 that's not it, C87294CE no. Ah! 452F6E403CDF10714E41DFAA257D313F.
Cook real food. One the range or in the oven. Disregard microwave ovens.
Nice, I know the second author of that since I just got my masters there in CS. Also, screw 'UW-Madison'. It is 'UW' and I hate that people put it on the same level as all the other bullshit UW-* schools (http://www.wisconsin.edu/campuses/).
First of all, switching channels to avoid a microwave is futile... the magnetron isn't all that frequency stable and the peak tends to wander across the band as a result.
Second, 802.11g/n uses OFDM. You get narrowband interference, you reduce the rate on the affected subcarriers. It's built in.
Third, I'm fairly sure using a wifi card as a spectrum analyzer has been done before.
This is cool but I would be far more interested in a card that does NOT detect microwave signals.
My 3 yr old dell D830 has a BCM4315/BCM22062000 chipset with a utility that detects congestion in the wireless G band. Used it to find an Epson printer that was committing a wireless DOS attack on my network. These things are nice to have if you are a geek for hire. Unfortunatly, the bastrds at Apple pulled all useful apps like this for their walled garden because the script kiddies were using them for malfesance. Bummer.
In this ideal world of yours, what incentive is there for ISPs to maintain their networks, as you have effectively cut them out of the loop?
You do realize that you would need a backbone SOMEWHERE unless you wanted horrific latency, for example to get traffic from coast to coast?
In 1992 I was at an IEEE 802.11 meeting (that's WiFi, if you didn't know it by that name). My company was presenting a "pre-standard" wireless LAN design that we were developing, to be considered as a contribution to the standard.
Someone asked "Why does your design have so much error correction coding? Are you expecting the RF environment to be that bad?"
I replied, "Well, I haven't seen any 'Listen Before Cook' microwave ovens out there!"
This got a few chuckles and we moved along.
Many years later, I was doing some patent searches, and I came upon Patent number 6,346,692, titled "Adaptive Microwave Oven"
I'll be damned! Somebody actually patented the "Listen Before Cook microwave oven!"
So now we have WiFi devices detecting microwave ovens. That seems obvious to me. But I'm still waiting for a commercially available microwave oven that will avoid stomping on my WiFi signal :-)
FWIW, The 802.11 Media Access Control (MAC) protocol effectively avoids microwave ovens most of the time, because the magnetrons in consumer microwave ovens only operate on a "half wave" basis. This means they're off at least half the time. A microwave oven during its "on" time looks indistinguishable from another WiFi transmitter, and so your WiFi device simply waits until the microwave oven turns off before transmitting the next packet. This results in slower throughput, but isn't a show stopper.
The bigger problem is that since the microwave oven doesn't listen before turning on its magnetron, it tends to "stomp on" your WiFi signal occasionally. This, combined with the fact that the majority of IP based communications is TCP (and TCP sees every packet loss as congestion, causing it to slow down for the next few-to-tens of seconds), results in more throughput loss than is strictly related to the number of packets "stomped upon."
Google already owns Meraki, builders of great mesh gear that does just this. There are multitudes of open source projects also with low cost gear. San Francisco is meshed nicely if you look at their map at meraki's website. Of course you do need gateways, 5 hops max before latency becomes a real issue, and how do you stop the freeloaders? In a perfect world it'd be nice, realistically it's a neat concept but a niche market.
I could so see an application that let you detect and pinpoint a phone or any bluetooth device that you lost around the house. The problem is a thief could also be using it to pinpoint what houses got the good electronic shit just by driving around or what houses are empty just by looking for who not using anything! Hell you want to steal dozen of cell phones you just drive to the nearest parking lot with this thing.
YOU do realize that if we want or need that backbone, which we do, as you correctly point out, then we will motherfucking pay for having that backbone. No reason to pay for unnecessary shit. Fucking corporate asslicker idiot.
I recently bought a wispy plus for about $200.
It was worth every cent
Perhaps the author lives in a cardboard box
Drat
Watch this Heartland Institute video
say 15' away, having a device cranking out enough power to go 200' is way too much power
The WiFi routers I've seen have a setup were one can adjust the output power. Of course, almost noone sets it to something sensible, which means my neighbour's signals a stronger than my own in parts of the apartment (on the other hand, that same inability to configure WiFi means they all fight around the channels 1-4, so I can avoid them).
Routers should have an automatic power adjustment for the devices it has connected. Is this technically possible?
LTE does something like this, where the base station continuously analyzes the radio spectrum and
figures out the optimal frequency (it uses multiple separate ones for the same cell) and the optimal
transmission power (for both downlink and uplink transmissions).
Part of the definition of optimal here is not so strong as to interfere with neighbouring cells more than necessary
In theory, you could implement some inter-AP protocol as part of a WiFi standard to allow them to
determine their resepective interference patterns on each other and cooperate better.
My magnetron is fed a sample from my wifi transmitter so nuking food doesn't stop me from nuking Zergs. :O
Yup, Shannon be damned. Frequency agility, which has been used in various radios for decades, will solve all these pesky spectrum crowding problems.
They should have done a little Googling before they decided on a name...there's already a commercial product (albeit a very different one) that uses the name "AirShark":
http://www.itrbo.com/airshark/airshark.html
For your security, this post has been encrypted with ROT-13, twice.
IIRC there a program back in the ethereal/early wireshark days called airshark that was for analysing WiFi traffic. A google & bing search turns up nothing. Anyone else remember this?
Pretty cool stuff indeed! I was wondering if anyone can comment on the utility of this new feature. So one can detect that a microwave oven or Xbox is operating nearby. I am thinking of war-driving for Xboxes now ;-) What else?
Doesn't Cisco clean air already do this? I wonder if they use a WiFi card or they use some special hardware built into the AP to detect the interference from these other devices?
Now what we need is a detector that can detect Barak Obama's pinus within 50 m.
This IS highly needed to protect cubicle workers at the "WhiteHouse" and other Government installlations in down town D.C. and neighborhoods in southeast D.C. from a Obama Pinus attack.
Nut Bunkers could be set up using sandbags and 50 caliber machine gun to foil the Pinus attack.
Marine detachment with flamthrower stand ready to burn the Pinus to a cinder.
SAC, the Statigic Air Command stands at the ready to lauch nuclear counter attacks on US Cities to foil the Obama Pinus attack,
Health Officals and Doctor at the CDC, Center of Disease Control stand in the breach to foil deadly desease from Obama Pinus attack and to give away free condoms to Blacks, Cubans and African-American. What a relief. Keep up the good work boys.
All this good will while the President, Barak Hussein Obama II plots his next fendish move to poison billions and otherwise disturpt the livelyhoods of peoples around the world for his selfish aims.
--
Yes, which is why I added "so much".
They should have done a little Googling before they decided on a name...there's already a commercial product (albeit a very different one) that uses the name "AirShark":
http://www.itrbo.com/airshark/airshark.html
Duh - same product.
Let me get this straight. These researchers managed to prove, on what was likely a proprietary platform, what we already proved on an open platform? The Ubertooth has been proving and showing this interference for months now. [Hak5 - Youtube]
The WiFi devices actually cooperate, it's part of the spec. A WiFi network consists of more than just the clients attached to it - it includes *all* clients and APs on the same channel as well.
First, the protocol is CSMA/CA, so any device using channel 1 (regardless of network) will try to avoid stomping on another device using channel 1, even if it's on a completely different (e.g., your neighbour's) network. It also has forward compatibility built in by setting a virtual carrier timeout - it's how 802.11n devices avoid conflicting with 802.11b/g devices. This virtual carrier basically tells incompatible devices that the medium is busy (for the specified duration) even if they can't decode the data.
And while you can DoS a network by doing this, you basicaly shut down the entire channel as everyone obeys it.
It's one of the reasons why setting a network to "g only" didn't really work (it was ignored if they detect a b device, which can be operating on another network, or in ad-hoc mode).
Interference still occurs though because you have classic radio problems of hidden nodes and such. But the IEEE recognized that nodes interfering with each other gave everyone less bandwidth than if they cooperated.
Fucking corporate asslicker idiot.
Oh yes, I see. Clearly it is unacceptable to have most of the work I do, as well as most of my clients, rely on sub-1000ms latency getting to the internet.
Honestly, who do you know who would put up with 1second skype latency, or 25 seconds to load any modern webpage (which will ping several domains, as well as do AJAXy requests, all incurring massive latency on a mesh network)? Do you have any idea just how bad the internet would be if everything you do incurred large fractions of a second in latency? Do you understand how awful slashdot would be on such a network?
Hmm....
the actual wifi implementations lack a lot of features. First, automation. No more configuration. Please.
- the channel should adjust automatically
- the same for the TX power, which is non trivial, due to pumping effects (one TX increases, the next one is disturbed, increases, the first one increases again, etc.)
- there should be no distinction between AP and normal node. A link administrative node should be selected automatically and dynamically
- use cooperative. each node as an IPV6, let the high levels do the routing. NOT the radio link. Relaying is then pretty straightforward. The performance lost in the bigger layer traversa will be gained by a more optimal network routing.
- each node sends a beacons for each other to listen, every 2-4 sec. KISS for network detection. optimal power use
- Please: forbid the actual policy to use an open wifi with a fake DNS opening a login window. Forbids any automation. Total ISP bullshit ! Forbidding iPV4 ant NATS would also resolve the ISP -made problems of wifi (compartmentalization), as well as fears of someone on one's local network (no NAT in IPv6, no local zone...)
- Please: mandatory routing. A node with an unlimited internet acess should provide it to the network. Access policy ? easy : give prio to the own served nodes, share evenly the (eventual) remaining BW.
- Crypto : not on the radio layers. Should be over IP. Please. Solutions exist. If this is properly implemented, it could be used for the rest of the internet as well. I'm not a specialist. but solutions exist : http://events.ccc.de/congress/2010/Fahrplan/events/4295.en.html Crypto has nothing to do at a local link level.
Please. Please. cooperative. People have enough of inefficient & closed networks.
Keep it simple. Darn simple. optimizing a radio link is complex enough (read some 3GPP specs, it's 10k+ pages, and that's centralized,not ad-hoc !!) so that routing, crypto, etc should not be on the radio levels !
Break compatibility. drop the actual BS. All at once. Call it a new standard. OEMS will follow.
F4eru at free.fr
Let's consider a cheap-made oven.
The door does not close so well. Door surfaces do not at all contact electrically, only an undefined capacitive coupling is taking place. The front grill has holes much smaller than the wavelength, but still leaks a negligible bit of power. Filters on the magnetron are basic LCs main cable couples that out....
Let's say you are lucky get a shielding efficiency of 40 dB, which is OK for such poorly constructed shields.
Inside this shield you have a hell of standing waves, with crazy fields, hot spots, and kilowatt of power being reflected back and forth leaky walls. Each reflection, -40 dB couples out.That's 1/10 000 of 1kW => 0.1W Suppose there are 10 reflections in average before absorbtion. The leak is then in the ~1W which is in the same order than a Wifi Transmitter.