4 ms·
Practical attacks against 4G (LTE) access network protocols
- chinathrow 11y agoOn a related note, Google still hasn't added TLS to blogspot.com...
- aiiane 11y agoIt's available to be turned on for individual blogs: https://support.google.com/blogger/answer/6284029?hl=en https://support.google.com/blogger/answer/6284029?hl=en
- ck2 11y agoThe LTE hack when it was first announced made me a little ill. http://www.zdnet.com/article/at-t-t-mobile-verizon-vulnerable-to-several-lte-flaws/ http://www.zdnet.com/article/at-t-t-mobile-verizon-vulnerabl... I always avoid wifi thinking cell data is safer but apparently not at all. Sometimes I wish phones/tablets had ethernet ports.
- dogma1138 11y agoIf you are connected to an untrusted (public) network it Ethernet vs. WiFi makes very little difference.
- ck2 11y agoYes but even on your own network, you can't trust wifi, unless you live in a Faraday cage.
- dogma1138 11y agoReally depends on the network you setup, yes the traffic can be picked over the air but if you secure the network sufficiently and only trusted devices are connected to it is trustworthy. People will not be easily breaking WPA2, if you rotate your keys, and implement other security measures such as RAPD or .x1 your threat agents change from anyone with a cellphone to very well funded and "armed" adversaries.
- simoncion 11y ago> ...but even on your own network, you can't trust wifi... WPA2 Enterprise uses AES for per-client session key generation, and -when you use EAP-TLS or EAP-TTLS- TLS for transportation of client credentials and assurance that both ends of the conversation are who they say they are. [0] You can make WiFi just as secure as wired Ethernet, it just takes a -surprisingly small- bit of work. :) If Windows and Mac machines supported EAP-PWD, or (even better) WFA-TLS-UNAUTH [1] even coffeeshop WiFi would be secure from passive attack. [0] When you use WiFi client and RADIUS server certs, rather than just username and password. [1] EAP offers a very large array of authentication methods. EAP-PWD authenticates with just a username and password and doesn't muck about with TLS certs. WFA-TLS-UNAUTH uses TLS, but doesn't verify the RADIUS server cert and doesn't require a client cert. Both of these schemes offer substantially less assurance than full-blown EAP-[T]TLS, but they also let you offer encrypted WiFi without hassle.
- xorcist 11y agoWhat's "safer" (or less safe) about different transport protocols, when you're carrying Internet traffic?
- bigiain 11y agoIt take a somewhat more privileged network position to MITM a 4G connection from your phone than "some guy with a wifi pineapple" impersonating some wifi base station your phone already trusts. It's easy enough to set up a wifi radio which says: "You wanna connect to 'Netgear'? Or 'Starbucks Free Wifi'? Sure - I can be that for you!". At least if I'm using 4G at Starbucks you'd need to be sitting on a backbone or in front of redtube^h^h^h^h^h^h^h wikipedia.com's servers to MITM me...
- bigiain 11y agoPossibly useful factoid - an iPad with a USB->Ethernet adaptor plugged into the usb port adaptor from the camera connection kit happily detects the internet connection and negotiates a DHCP lease. (Caveat, I haven't tried this since maybe mid 2013 or so...)
- rmac 11y agotldr: 1) LTE uses broadcast messages to do push notifications. These are unencrypted and unauthenticated. Researchers can trigger them silently by spamming you on Facebook, and triangulate your location since the LTE network only sends the pages to the geographic location you are in. 2) LTE supports crash reports that can potentially include your GPS coords. A rouge base station can request these from your phone without setting up an encrypted channel due to shitty baseband implementations. 3) something about denial of service the excellent paper is here: http://arxiv.org/pdf/1510.07563v1.pdf http://arxiv.org/pdf/1510.07563v1.pdf
- TorKlingberg 11y agoI'd like to add a few things: 1) This is really a fundamental limitation of all encryption: It hides what you are sending, but not the fact that you are sending something. Paging in mobile networks is a broadcast message that basically just says "Device with ID XXXX, please wake up and check in with the network". If you can cause pages (by calling, texting or with Facebook messages) then you can see in what area the pageing is sent, and you can find out the ID (TMSI) of the device. 2) It is not a crash report, but a connection failure report. 3) During the initial connection, before encryption is started, an rogue transmitter can send a connection rejection message to the phone, pretending to be the network. The phone will not try 4G/3G again and downgrade to 2G. The attacker can then break the weak 2G authentication and make the phone believe that it is a real base station. The last part is what an IMSI catcher / Stingray device does.
- Natanael_L 11y ago1) but you can hide who it is sent to, such that only timing attacks are plausible (see Tor and every VPN)
- zurn 11y agoPaper says they are leveraging GNU Radio hardware (USRP) and using parts of this SDR LTE implementation: https://github.com/srsLTE/srsLTE https://github.com/srsLTE/srsLTE Pretty cool. I wonder how much of "some assembly required" there is to get this kind of experiment up and running.