I found this bit about hazards uniquely faced by internation travelers very interesting: Q: Do you think organizations are understanding and taking the threat among these new mobile attack vectors seriously yet? Are security managers really getting it? Why or why not? A: The most-security-aware organizations are taking these threats very seriously. They're destroying phones after taking them to hostile areas with known malicious carriers, they're limiting what information gets copied to the default inbox/contact list on devices, they're limiting what applications can be installed on devices which have access to enterprise infrastructure. [...] Q: In your presentation, you specifically referred to some of the threats mobile users are facing now while traveling internationally. What are you observing? A: There are two major threat categories when it comes to international travel, the malicious foreign carrier and the enterprising private mobile attacker. These threats result from the fact that citizens of a foreign country generally have no rights to privacy and no official recourse if their information gets stolen while they are in the foreign country. I already spoke about how foreign carriers have total control over devices which are associated with their networks. Probably the most alarming thing we've seen happen in our tests is how foreign carriers can steal the cryptographic seed values from soft-tokens installed on smartphones. One take-away I'd love to get across to all of your readers is to never let soft-tokens become a solution to be relied on for organizations which have a large number of international travelers. [...] http://www.csoonline.com/article/print/733691 I'm wary of distortions, outright bullshit, and vague handwaving by security industry guys. As usual, interviewee Aaron Turner (of IntegriCell, formerly Idaho National Laboratory, formerly Microsoft security division) is vague about how code gets executed to do undesirable things. Smartphone 'soft-tokens' means virtual dongles used for (supposed) two-factor authentication such as Symantec VIP or SecurEnvoy SecurAccess. Pity Turner doesn't say _how_ iOS, Android, and BlackberryOS yield 'total control over devices' to local carriers. Turner goes on, in the bit about international travelers, to talk about a cruise ship passenger who uses a cafe's WiFi in port, and 'the coffee shop owner has realized he can make more money selling your address book to spearfishers than he ever can make selling you even his most-expensive latte'. I'm intrigued, Mr. Turner. How does connecting a smartphone to a WAP give the WAP owner automatic access to the smartphone's address book? I'm used to assuming that networks are dangerous and untrustworthy and layering things trustworthy over them. Why is that approach not feasible with a smartphone -- or is Turner just another guy selling unnecessary gear to the rubes?
On Sat, 27 Jul 2013, Rick Moen <rick@linuxmafia.com> wrote:
A: The most-security-aware organizations are taking these threats very seriously. They're destroying phones after taking them to hostile areas with known malicious carriers, they're limiting what information gets copied to the default inbox/contact list on devices, they're limiting what applications can be installed on devices which have access to enterprise infrastructure. [...]
Which would be more of a risk? A random drive-by attack against all users of a rogue telco in some strange country or a random drive-by attack with a fake cell tower in the center of the Melbourne CBD or other well trafficed area? I presume the latter if compromising phones via mobile tower is so easy.
A: There are two major threat categories when it comes to international travel, the malicious foreign carrier and the enterprising private mobile attacker. These threats result from the fact that citizens of a foreign country generally have no rights to privacy and no official recourse if their information gets stolen while they are in the foreign country. I already spoke about how foreign carriers have total control over devices which are associated with their networks. Probably the most alarming thing we've seen happen in our tests is how foreign carriers can steal the cryptographic seed values from soft-tokens installed on smartphones. One take-away I'd love to get across to all of your readers is to never let soft-tokens become a solution to be relied on for organizations which have a large number of international travelers. [...]
Given that telcos in countries like the US have deliberately sold phones with malware pre-loaded, I expect that such things would be done in countries with a supposed rule of law too if they were so easy. Of course the rule of law in the US only applies to poor people.
Smartphone 'soft-tokens' means virtual dongles used for (supposed) two-factor authentication such as Symantec VIP or SecurEnvoy SecurAccess. Pity Turner doesn't say _how_ iOS, Android, and BlackberryOS yield 'total control over devices' to local carriers.
It seems that every time a new version of iOS is released someone jail-breaks it quite quickly. So it's obviously not impossible to crack the OS. But doing so reliably over a population without getting caught is going to be difficult.
Turner goes on, in the bit about international travelers, to talk about a cruise ship passenger who uses a cafe's WiFi in port, and 'the coffee shop owner has realized he can make more money selling your address book to spearfishers than he ever can make selling you even his most-expensive latte'. I'm intrigued, Mr. Turner. How does connecting a smartphone to a WAP give the WAP owner automatic access to the smartphone's address book?
There's also the economic issue. Why would someone who wants such data buy from small cafes? If it was so easy to do then someone could setup wifi acces points purporting to be McDonalds, Starbucks, and other common vendors in central city areas. Phones usually connect to such Wifi access points by default without even informing the user so if such a Wifi attack is possible then it could be implemented against 10s of thousands of people a day in Melbourne CBD instead of dozens a day in a backwater cafe.
I'm used to assuming that networks are dangerous and untrustworthy and layering things trustworthy over them. Why is that approach not feasible with a smartphone -- or is Turner just another guy selling unnecessary gear to the rubes?
Sounds like it. It seems to me that one of the biggest problems with security is the use of devices for multiple purposes. If a company issues a nice laptop and phone to an employee then there's a good chance the laptop will be used to view porn and the phone will be used for playing games (particularly if the employee has children) and that's two major vectors for attack. One of my suggestions has been to issue laptops with small screens to make them less suitable for porn. For phone use when employees access sensitive data it wouldn't be particularly expensive to give them a spare phone for games. -- My Main Blog http://etbe.coker.com.au/ My Documents Blog http://doc.coker.com.au/
On 28 July 2013 11:19, Russell Coker <russell@coker.com.au> wrote:
Smartphone 'soft-tokens' means virtual dongles used for (supposed) two-factor authentication such as Symantec VIP or SecurEnvoy SecurAccess. Pity Turner doesn't say _how_ iOS, Android, and BlackberryOS yield 'total control over devices' to local carriers.
It seems that every time a new version of iOS is released someone jail-breaks it quite quickly. So it's obviously not impossible to crack the OS. But doing so reliably over a population without getting caught is going to be difficult.
Just because it is possible to jail-break the phone, which usually requires physical procession of the phone, does not mean your carrier can remotely access your data. In fact, at least for some phones jail breaking can't be done without erasing your data, so they still wouldn't be able to access it. Maybe the article is talking about bugs/security exploits in the OS that may or may not be fixed in recent code? A good reason not to use old versions of Android. I would hope everyone is using secure encryption these days for syncing contacts, etc. However, if not, that is another way your carrier could get access to your data. -- Brian May <brian@microcomaustralia.com.au>
Rick Moen wrote:
A: The most-security-aware organizations are taking these threats very seriously. They're destroying phones after taking them to hostile areas with known malicious carriers
IIRC for the beijing olympic games, BHP gave their people[0] shiny new thinkpads to take to China, instead of taking their existing corporate laptops. I heard ASIO (WTF? Not ASIS or DFAT?) told them to expect Chinese intelligence to take laptops out of hotel safes, dd them, then put them back. I don't think they do it for normal visits to e.g. Surinam and South Africa, tho. [0] though probably only management, not the plebes.
On 29 July 2013 10:30, Trent W. Buck <trentbuck@gmail.com> wrote:
Rick Moen wrote:
A: The most-security-aware organizations are taking these threats very seriously. They're destroying phones after taking them to hostile areas with known malicious carriers
IIRC for the beijing olympic games, BHP gave their people[0] shiny new thinkpads to take to China, instead of taking their existing corporate laptops. I heard ASIO (WTF? Not ASIS or DFAT?) told them to expect Chinese intelligence to take laptops out of hotel safes, dd them, then put them back. I don't think they do it for normal visits to e.g. Surinam and South Africa, tho.
Also think of: http://xkcd.com/538/ You might think you are protected by having all your files encrypted, then the foreign country demands you give them your decryption key, with threats of jail if you don't. I wouldn't take any sensitive information on computer/phone/tablet overseas, even encrypted, unless I absolutely had to. Also consider that a hostile entity might be able to get cached browser information, find out what Google, Twitter, Facebook, etc, accounts you have, then coerce you to log into them. -- Brian May <brian@microcomaustralia.com.au>
On 29 July 2013 12:19, Brian May <brian@microcomaustralia.com.au> wrote:
Also consider that a hostile entity might be able to get cached browser information, find out what Google, Twitter, Facebook, etc, accounts you have, then coerce you to log into them.
... assuming of course you haven't saved your login or credentials unencrypted in the browser, in which case it becomes even easier. -- Brian May <brian@microcomaustralia.com.au>
Quoting Brian May (brian@microcomaustralia.com.au):
You might think you are protected by having all your files encrypted, then the foreign country demands you give them your decryption key, with threats of jail if you don't.
Schneier has had a number of articles about creative ideas to deal with that scenario. There are some very creative people out there, and I cannot remember most of the nuances. One was a filesystem that, if you give it one key gives access to the real data but if given a different key gives access to innocuous decoy data. I've also considered storing on any traveling laptopc some large files containing random data and publicising widely the fact that I do so.
I wouldn't take any sensitive information on computer/phone/tablet overseas, even encrypted, unless I absolutely had to.
Yes, it's an interesting problem. One possibility is to bring only a generically loaded machine with the intent of later wiping it upon reaching one's destination and downloading the intended substantive contents off an Internet repository established in advance for that purpose. The trick would be to keep the download size and transfer time reasonable.
Rick Moen wrote:
Schneier has had a number of articles about creative ideas to deal with that scenario. There are some very creative people out there, and I cannot remember most of the nuances. One was a filesystem that, if you give it one key gives access to the real data but if given a different key gives access to innocuous decoy data.
Assange had a patch to ext ages ago (like, when ext2 was current) to give you multiple filesytems inside ext, with deniability. The game theory goes something like "unless ALL blocks are allocated to filesystems you know about, you CAN'T know that I've given you all the access keys, therefore you have no reason to ever stop hitting me with that rubber hose". I'm not sure how that works in practice. If you just mounted it as normal ext, it'd consider those secret filesystems' blocks as not reserved, so they might be trashed by writes.
Yes, it's an interesting problem. One possibility is to bring only a generically loaded machine with the intent of later wiping it upon reaching one's destination and downloading the intended substantive contents off an Internet repository established in advance for that purpose. The trick would be to keep the download size and transfer time reasonable.
And avoiding MITM if $bad_guy can lift the key material off your person / out of your head beforehand.
On 2013-07-29 18:19, Trent W. Buck wrote:
Rick Moen wrote:
Schneier has had a number of articles about creative ideas to deal with that scenario. There are some very creative people out there, and I cannot remember most of the nuances. One was a filesystem that, if you give it one key gives access to the real data but if given a different key gives access to innocuous decoy data.
Assange had a patch to ext ages ago (like, when ext2 was current) to give you multiple filesytems inside ext, with deniability.
The game theory goes something like "unless ALL blocks are allocated to filesystems you know about, you CAN'T know that I've given you all the access keys, therefore you have no reason to ever stop hitting me with that rubber hose". I'm not sure how that works in practice.
If you just mounted it as normal ext, it'd consider those secret filesystems' blocks as not reserved, so they might be trashed by writes.
For reference, the filesystem was called Marutukku or Rubberhose: http://en.wikipedia.org/wiki/Rubberhose_%28file_system%29 https://github.com/sporkexec/rubberhose -- Regards, Matthew Cengia
On 29/07/2013 7:55 PM, Matthew Cengia wrote:
For reference, the filesystem was called Marutukku or Rubberhose:
Truecrypt can do similar, you can have multiple containers and depending on the key used, different containers will open up. Cheers A.
On 29/07/2013 7:55 PM, Matthew Cengia wrote:
For reference, the filesystem was called Marutukku or Rubberhose:
Truecrypt can do similar, you can have multiple containers and depending on the key used, different containers will open up. http://www.truecrypt.org/docs/plausible-deniability Cheers A.
Brian May wrote:
IIRC for the beijing olympic games, BHP gave their people[0] shiny new thinkpads to take to China, instead of taking their existing corporate laptops. I heard ASIO (WTF? Not ASIS or DFAT?) told them to expect Chinese intelligence to take laptops out of hotel safes, dd them, then put them back. I don't think they do it for normal visits to e.g. Surinam and South Africa, tho.
[...] I wouldn't take any sensitive information on computer/phone/tablet overseas, even encrypted, unless I absolutely had to.
In case that's not obvious, that was (AIUI) BHP's point. You're going there to have fun, not do deals, so your corporate secrets stay at home on your "real" work laptop. This one is just for skyping to your family or whatever normal people do.
Quoting Trent W. Buck (trentbuck@gmail.com):
IIRC for the beijing olympic games, BHP gave their people[0] shiny new thinkpads to take to China, instead of taking their existing corporate laptops. I heard ASIO (WTF? Not ASIS or DFAT?) told them to expect Chinese intelligence to take laptops out of hotel safes, dd them, then put them back. I don't think they do it for normal visits to e.g. Surinam and South Africa, tho.
Sounds about right. Back in 1999, I worked as head of the system administration department at a Linux-industry pre-IPO firm[1] where I had an excellent hunch, eventually proved correct, that the new CTO was somehow tapping all company electronic communications. (Later, it emerged that he was monitoring via the ethernet switches.) I reasoned that, if I couldn't trust the integrity of company infrastructure generally, I also couldn't fully trust my Debian-based personal workstation, either, as he might be able to either trojan it subtly enough to make detection difficult or install a hardware keylogger. Because the notion of this particular person having any form of visibiltiy into my _personal_ activity between me and my server at home, I bought for the first time a laptop, always used it rather than my workstation for anything sensitive, and never, ever left it unattended. [1] http://linuxmafia.com/~rick/essays/preipo.html
Rick Moen wrote:
Because the notion of this particular person having any form of visibiltiy into my _personal_ activity between me and my server at home, I bought for the first time a laptop, always used it rather than my workstation for anything sensitive, and never, ever left it unattended.
And never used an external CRT (ref. TEMPEST), and... :-)
On 07/27/2013 06:21 AM, Rick Moen wrote:
I'm used to assuming that networks are dangerous and untrustworthy and layering things trustworthy over them. Why is that approach not feasible with a smartphone
Depends on the phone but mostly the proprietary baseband firmware (and other blobz) and the modems direct memory access = your phone is pwned by any serious adversary be them international or local. OTA updates can be pushed silently by a telco by design even on a non smart phone. they simply are not secure.
On 29 July 2013 12:25, nic <nic@404ed.org> wrote:
OTA updates can be pushed silently by a telco by design even on a non smart phone. they simply are not secure.
Only if you get OTA updates from your telco. Even then I would assume it will only get updates from the telco that provided you with the phone. Which could be different from the telco that you are currently connected to, especially if travelling overseas. Doesn't apply if you didn't get your phone from a telco or installed a custom version of Android on it. -- Brian May <brian@microcomaustralia.com.au>
On 29/07/2013 1:18 PM, Brian May wrote:
On 29 July 2013 12:25, nic <nic@404ed.org <mailto:nic@404ed.org>> wrote:
OTA updates can be pushed silently by a telco by design even on a non smart phone. they simply are not secure. Only if you get OTA updates from your telco. Even then I would assume it will only get updates from the telco that provided you with the phone. Which could be different from the telco that you are currently connected to, especially if travelling overseas.
Doesn't apply if you didn't get your phone from a telco or installed a custom version of Android on it.
My S3 gets OTA updates, I bought it outright and unbranded -- no carrier owns my phone; however, I'm sure that they could if they wanted to given how easy it is to push OTA updates. Cheers A.
Andrew McGlashan <andrew.mcglashan@affinityvision.com.au> wrote:
My S3 gets OTA updates, I bought it outright and unbranded -- no carrier owns my phone; however, I'm sure that they could if they wanted to given how easy it is to push OTA updates.
Isn't there a signature verification involved? For example, if I were getting updates from a phone manufacturer rather than a carrier, I would expect the phone to hold the manufacturer's public key and accept only updates signed by the corresponding secret key.
On 07/29/2013 01:18 PM, Brian May wrote:
Only if you get OTA updates from your telco. Even then I would assume it will only get updates from the telco that provided you with the phone. Which could be different from the telco that you are currently connected to, especially if travelling overseas.
Doesn't apply if you didn't get your phone from a telco or installed a custom version of Android on it.
any carrier you connect to via GSM can send push OTA to your phone/sim its not an android update but a FTOA update of one of the proprietary blobs on your handset once they have that level access its game over. hence the problematic nature of the closed source baseband drivers and direct memory access given to the modem on most phones.
On 2013-07-29 13:44, nic wrote: [...]
any carrier you connect to via GSM can send push OTA to your phone/sim its not an android update but a FTOA update of one of the proprietary blobs on your handset once they have that level access its game over. hence the problematic nature of the closed source baseband drivers and direct memory access given to the modem on most phones.
Nic, Could you please make a more concerted effort to use full sentences with proper grammar and punctuation? Your first sentence above is very difficult to understand. I've not investigated this sort of update before, but based on the little I grasped from your email I'd be interested to learn more about how, when, and why these updates are deployed. -- Regards, Matthew Cengia
On 07/29/2013 02:10 PM, Matthew Cengia wrote:
On 2013-07-29 13:44, nic wrote: [...]
any carrier you connect to via GSM can send push OTA to your phone/sim its not an android update but a FTOA update of one of the proprietary blobs on your handset once they have that level access its game over. hence the problematic nature of the closed source baseband drivers and direct memory access given to the modem on most phones.
Nic,
Could you please make a more concerted effort to use full sentences with proper grammar and punctuation? Your first sentence above is very difficult to understand.
I shall do my best.
I've not investigated this sort of update before, but based on the little I grasped from your email I'd be interested to learn more about how, when, and why these updates are deployed.
http://en.wikipedia.org/wiki/FOTA_(technology) http://en.wikipedia.org/wiki/FUMO http://en.wikipedia.org/wiki/OMA_DM The above is ok for a very basic overview of the updating process allowing updating of the baseband via a number of ways. Below is focussed on exploiting the baseband rather than a malicious update its interesting given this was a non-nation state attacker, so they did not have the advantage of a "relationship" with a carrier/operator https://www.usenix.org/system/files/conference/woot12/woot12-final24.pdf
On 29 July 2013 15:31, nic <nic@404ed.org> wrote:
I've not investigated this sort of update before, but based on the little I grasped from your email I'd be interested to learn more about how, when, and why these updates are deployed.
http://en.wikipedia.org/wiki/FOTA_(technology) http://en.wikipedia.org/wiki/FUMO http://en.wikipedia.org/wiki/OMA_DM The above is ok for a very basic overview of the updating process allowing updating of the baseband via a number of ways.
Below is focussed on exploiting the baseband rather than a malicious update its interesting given this was a non-nation state attacker, so they did not have the advantage of a "relationship" with a carrier/operator
https://www.usenix.org/system/files/conference/woot12/woot12-final24.pdf
Thanks for the links. According to http://en.wikipedia.org/wiki/OMA_DM, it only applies to certain phones. Only one really old Android phone appears in the list. Are you sure this is an issue? Yes, more companies are listed under http://en.wikipedia.org/wiki/FOTA_(technology), however I think this is a more generic thing, and probably applies to the Android updates. -- Brian May <brian@microcomaustralia.com.au>
Brian May <brian@microcomaustralia.com.au> wrote:
https://www.usenix.org/system/files/conference/woot12/woot12-final24.pdf
Thanks for the links.
According to http://en.wikipedia.org/wiki/OMA_DM, it only applies to certain phones. Only one really old Android phone appears in the list.
Note however that the attacks described in the paper cited above don't rely on the delivery of firmware updates; they exploit security vulnerabilities in the GSM implementation. Thus there is cause for concern (about the binary-only drivers and what might be lurking therein) quite aside from the firmware updating mechanism.
participants (8)
-
Andrew McGlashan -
Brian May -
Jason White -
Matthew Cengia -
nic -
Rick Moen -
Russell Coker -
Trent W. Buck