Linked by Thom Holwerda on Sat 19th Jul 2014 19:06 UTC
Apple

Jonathan Zdziarski's paper about backdoors, attack points and surveillance mechanisms built into iOS is quite, quite interesting.

recent revelations exposed the use (or abuse) of operating system features in the surveillance of targeted individuals by the National Security Agency (NSA), of whom some subjects appear to be American citizens. This paper identifies the most probable techniques that were used, based on the descriptions provided by the media, and today’s possible techniques that could be exploited in the future, based on what may be back doors, bypass switches, general weaknesses, or surveillance mechanisms intended for enterprise use in current release versions of iOS. More importantly, I will identify several services and mechanisms that can be abused by a government agency or malicious party to extract intelligence on a subject, including services that may in fact be back doors introduced by the manufacturer. A number of techniques will also be examined in order to harden the operating system against attempted espionage, including counter-forensics techniques.

This paper is actually half a year old - give or take - but it's gotten a lot of attention recently due to, well, the fact that he has uploaded a PowerPoint from a talk about these matters, which is obviously a little bit more accessible than a proper scientific journal article.

For instance, despite Apple's claims of not being able to read your encrypted iMessages, there's this:

In October 2013, Quarkslab exposed design flaws in Apple's iMessage protocol demonstrating that Apple does, despite its vehement denial, have the technical capability to intercept private iMessage traffic if they so desired, or were coerced to under a court order. The iMessage protocol is touted to use end-to-end encryption, however Quarkslab revealed in their research that the asymmetric keys generated to perform this encryption are exchanged through key directory servers centrally managed by Apple, which allow for substitute keys to be injected to allow eavesdropping to be performed. Similarly, the group revealed that certificate pinning, a very common and easy-to-implement certificate chain security mechanism, was not implemented in iMessage, potentially allowing malicious parties to perform MiTM attacks against iMessage in the same fashion.

There are also several services in iOS that facilitate organisations like the NSA, yet these features have no reason to be there. They are not referenced by any (known) Apple software, do not require developer mode (so they're not debugging tools or anything), and are available on every single iOS device.

One example of these services is a packet sniffer, com.apple.pcapd, which "dumps network traffic and HTTP request/response data traveling into and out of the device" and "can be targeted via WiFi for remote monitoring". It runs on every iOS device. Then there's com.apple.mobile.file_relay, which "completely bypasses Apple’s backup encryption for end-user security", "has evolved considerably, even in iOS 7, to expose much personal data", and is "very intentionally placed and intended to dump data from the device by request".

This second one, especially, only gave relatively limited access in iOS 2.x, but in iOS 7 has grown to give access to pretty much everything, down to "a complete metadata disk sparseimage of the iOS file system, sans actual content", meaning time stamps, file names, names of all installed applications and their documents, configured email accounts, and lot more. As you can see, the exposed information goes quite deep.

Apple is a company that continuously claims it cares about security and your privacy, but yet they actively make it easy to get to all your personal data. There's a massive contradiction between Apple's marketing fluff on the one hand, and the reality of the access iOS provides to your personal data on the other - down to outright lies about Apple not being able to read your iMessages.

Those of us who aren't corporate cheerleaders are not surprised by this in the slightest - Apple, Microsoft, Google, they're all the same - but I still encounter people online every day who seem to believe the marketing nonsense Apple puts out. People, it doesn't get much clearer than this: Apple does not care about your privacy any more or less than its competitors.

Thread beginning with comment 592787
To view parent comment, click here.
To read all comments associated with this story, please click here.
RE[3]: Darn
by Lennie on Sun 20th Jul 2014 12:45 UTC in reply to "RE[2]: Darn"
Lennie
Member since:
2007-09-22

Enrypted communications can solve that.

But it probably has access to other parts of the system too.

It is probably very similar to most servers where it has a Baseboard Management Controller (like HP Lights Out Management or Dell DRAC) for out-of-band management.

Which has direct access to the RAM used by the mainsystem.

The people from the neo900 project specifically mentions this in their FAQ:

Isn't a non-free baseband firmware a privacy issue?

We're going to address privacy concerns of non-free modem firmware by ensuring that modem has access to no other data than absolutely necessary, so it won't be able to spy on anything that's not already available on carrier side. On Neo900 one can be sure that the modem is actually turned off when requested, not just pretending to be. User will be notified in case of modem wanting to do something without his consent.

Unlike some other smartphones do, Neo900 won't share system RAM with the modem and system CPU will always have full control over the microphone signal sent to the modem. You can think of it as a USB dongle connected to the PC, with you in full control over the drivers, with a virtual LED to show any modem activity.

http://neo900.org/faq#privacy

Reply Parent Score: 5

RE[4]: Darn
by ssokolow on Mon 21st Jul 2014 08:18 in reply to "RE[3]: Darn"
ssokolow Member since:
2010-01-21


We're going to address privacy concerns of non-free modem firmware by ensuring that modem has access to no other data than absolutely necessary, so it won't be able to spy on anything that's not already available on carrier side. On Neo900 one can be sure that the modem is actually turned off when requested, not just pretending to be. User will be notified in case of modem wanting to do something without his consent.

Unlike some other smartphones do, Neo900 won't share system RAM with the modem and system CPU will always have full control over the microphone signal sent to the modem. You can think of it as a USB dongle connected to the PC, with you in full control over the drivers, with a virtual LED to show any modem activity.

http://neo900.org/faq#privacy


The DragonBox Pyra (pocketable Debian+ARM gaming laptop) is going to do something similar with its cellular modem.

1. The modem+GPS chip will be separate and you'll be able to buy units with or without it. (partly to save you money if you don't need it)

2. If you choose to have the modem, it connects to the rest of the system as an internal USB peripheral. No DMA.

3. It's not feasible to put a switch on the modem's power supply line to ensure a hack couldn't force it to ignore its enable/disable pin, so, instead, they're going to wire the cellular status LED up to the power controller as an "is the modem drawing power?" LED.

Reply Parent Score: 5

RE[5]: Darn
by Lennie on Mon 21st Jul 2014 08:57 in reply to "RE[4]: Darn"
Lennie Member since:
2007-09-22

thanks for the information

Reply Parent Score: 2