Netatmo Weather Station Sends WPA Passwords In the Clear 37
UnderAttack writes The SANS Internet Storm Center is writing that Netatmo weather stations will send the users WPA password in the clear back to Netatmo. Netatmo states that this is some forgotten debug code that was left in the device. Overall, the device doesn't bother with encryption, but sends all data, not just the password, in the clear. From the article: "After reporting the bug to Netatmo, the company responded, acknowledging that it does indeed dump all that data from the weather station’s memory unencrypted and that it would stop doing that the coming weeks."
Ahh, the internet of things... (Score:5, Interesting)
Re:Ahh, the internet of things... (Score:4, Informative)
What does this have to do with a newfangled marketing term? We've seen routers, access points and all manner of devices do this sort of thing since the 1990s - data leakage, deliberate or otherwise, its not a new thing.
Re:Ahh, the internet of things... (Score:5, Insightful)
Re:Ahh, the internet of things... (Score:4, Insightful)
Sometimes I think the situation is hopeless. Consumers want the lowest possible price and see many items as commodities. Security is expensive.
Re: (Score:3)
This.
While I don't give a shit if someone can read, in clear text, what the temperature is, or if they can even see the authentication, of a weather station, because it's a low-security device, the larger question is, "What is the threshold that separates innocuous devices from more serious devices?"
The "no security" feature of these IoT is a culture that can creep into the design of more important web-enabled devices.
It's a bad habit.
Re: (Score:3)
Honestly, this is just the on-going demonstration of the fact that most network-enabled consumer products are garbage, written by incompetent and lazy idiots, who neither know nor care about your security or privacy, and pushed out the door by greedy bastards.
Until there are real penalties for doing crap like this ... I just assume that all things which want to connect to the internet will probably be insecure and dangerous, and therefore won't trust them.
It's pretty much happening so often that it's a safe
Taste of things to come (Score:5, Insightful)
Wow that's a pretty big oversight. I work in hardware and this sort of stuff is pretty common. I worked for one medical device company that simply XORed their firmware with a fixed 8-bit value to 'encrypt' it. Trouble is that when the design team is trying to fix flow lines on plastic mouldings or get the product through 20k of EMC testing, software security falls to the bottom of the list, and typically a guy who knows how to write embedded code for reading sensors but has no idea what it really means to open a public facing port to the Internet.
One shudders to think what other debug back doors they have left in there and what sort of shonky TCP/IP library they found on the Internet to stuff into the firmware.
Re:Taste of things to come (Score:5, Interesting)
I'm looking forward to the first consumer protection law claim on a consumer IoT device. In the UK you could perhaps argue that the device is not fit for purpose, since it can't safely be connected to the internet. The shop you bought it from has "a reasonable length of time" to fix it, which typically means 28 days. If the manufacturer fails to provide an update in that time the shop is screwed and you can get a refund.
I'm hoping that kind of claim becomes more common. Someone in the UK already got a refund from Amazon when Sony removed features from the PS3. I read that Sony and a few others have already dropped YouTube support from some older smart TVs in Japan, and if it happens in the UK I'd be expecting a partial refund for loss of functionality. The formula is basically the amount of use you have had from the product vs. how long you would expect it to last, multiplied by how much functionality is lost. So, say I spend 1/3rd of my time watching YouTube (possibly an under-estimation, I have a lot of subscripts but don't watch that much TV overall):
3 year old TV, would reasonably expect it to last at least 10 years (5 year warranty, expensive plasma screen). So 66% of its reasonable life span remaining. 33% loss of functionality. Say I paid £1500 for this thing, I would expect a £326.70 refund if YouTube stopped working. Alternatively the shop could provide something else with equivalent functionality, such as a set top box or smart BluRay player.
It's a full memory dump (Score:5, Informative)
Netatmo states that this is some forgotten debug code that was left in the device.
It is actually a full memory dump which just happens to contain the WPA password. It seems to have been a legit debug feature, although it of course is a bit stupid that they have left it there. The quality assurance is still a bit crusty with these IoT devices.
Re:It's a full memory dump (Score:5, Interesting)
It is actually a full memory dump which just happens to contain the WPA password. It seems to have been a legit debug feature, although it of course is a bit stupid that they have left it there.
Yeeeeesss, very "stupid"
They "stupidly" just got themselves a map of APs and their passwords.
You're probably right, of course, but how could you distinguish this from an actual attack?
Re: (Score:2)
Because this would be a really stupid way to do an attack.
Re:It's a full memory dump (Score:4, Interesting)
Because this would be a really stupid way to do an attack.
Well, if you think it would be stupid, then it must be a really good way to do it.
The best thing about an attack done in this way is that the target doesn't know they were targeted. Since netatmo is so careless at security, they wouldn't even have had to have been the ones who made the attack. Someone else could have diddled their code and kept debug on in release.
Maybe I'm not keeping up? (Score:2)
I thought WPA was found to be insecure a long time ago! Are there really still entities that depend on it for security?
Re: (Score:2)
Are you sure? I thought WPA2 was created to solve the problems with WPA.
Re: (Score:2)
So WPA2 wasn't necessary? Really looking forward you to enlightening me here.
Re: (Score:1)
Re: (Score:1)
WPA uses rc4 encryption - which can be cracked by collecting enough packets encypted with the same key. WPA usually ensures the key is rotated before that happens though. WPA2 does not have this known weakness, so it considered to be better.
Re: (Score:3)
WPA uses rc4 encryption - which can be cracked by collecting enough packets encypted with the same key
Unless Im mistaken, RC4 is not in itself vulnerable or broken; it was used very widely in 2011 when AES was under siege by the BEAST attack (Google.com actually used it). The worst that could be said for it (as I understand) is that its a little too simple and fast for people to have full confidence in it, not to mention its age.
Based on my limited understanding of it, your statement about rotation is sort of kind of correct, but misleading in that it implies that the issue is with RC4 itself and not the
Re: (Score:2)
WPA uses portions from WEP, but AFAIK its not terribly vulnerable because it was designed better. There are "weaknesses" but as I recall current attacks on WPA basically boil down to bruteforce.
Re: (Score:3)
WPA is TKIP. It's a way of making the network more secure than WEP without requiring major hardware upgrades (uses the same RC4 cipher as WEP). WPA2 introduces AES encryption. IIRC there are known vulnerabilities that allow an attacker to inject packets into a WPA connection.
Re: (Score:1)
Re: (Score:2)
Correct, although many WiFi devices support WPA in AES mode too. WPA-AES with a strong preshared key should be pretty good.
Similarly, most devices also seem to support WPA2 in TKIP mode.
If you really want security, you must turn to cert-based IPSEC, with a cert per client. Yes, ugh. So much ugh. But otherwise, who knows what apps are stealing your credentials?
Re: (Score:1)
They *do* seem to overread compared with proper weather stations, if you look at wundermap - although that could be because they're sold more as a fashion accessory than a serious weather instrument and owners may not be siting them properly.
Phones send the same information to their vendors (Score:1)
So what else is new? People leave their phones to the default settings which makes them back the wifi passwords "to cloud". In practice the Apple and Google and their "partners" have access to millions of wifi networks. They just gather the data over SSL to avoid leaking that to competitors.