Security Weakness in P25

[rdale]

Anyone can write a paper. Post decoded audio files and then you'll come out of the whacker stage.

 

[kb0uxv]

If you are going to challenge someone to prove it, maybe you should provide the audio file to preserve the integrity of the test.

Let's assume that someone replies with a cracked audio file, or simply a transcript of the conversation in the voice file. Next comes questions of how, followed by a step by step instruction of how to do it yourself. Personally, I think that is dangerous and a bad idea.

Maybe we should just accept that it can be done, but only by those with very advanced knowledge and special equipment.

 

[rdale]

If you are going to challenge someone to prove it, maybe you should provide the audio file to preserve the integrity of the test.

Will do - he can PM me and I'll send him a .wav or any other format.

Maybe we should just accept that it can be done, but only by those with very advanced knowledge and special equipment.

I also wanted to mention that I flew to the moon and back yesterday. Even though it's not physically impossible, it sure is hard to believe. I have no evidence of it, so just assume that I have very advanced knowledge and special equipment

You can't come and make claims without SOME sort of evidence. A paper is cool. But I've seen WAY too many papers explaining theories that fell apart in the real world. That's all we're asking for if you want to start claiming you can decrypt P25 encryption on the fly.

 

[MattSR]

SecureComm 2011 7th International ICST Conference on Security and Privacy in Communications Network

Presentation 24.

You're welcome to fly to London and explain in front of everyone how and why we are wrong

 

[analogwarrior]

My all around opinion on project 25 is that it needs to go away! The more complex of a system you build, the more that can go wrong. A well built conventional system will do just fine. Emphasis should be on hardware based systems not complex massive software/firmware based ones. Other than to show how big their wallet is,why does a county/state/ect need to put an elaborate digital system in place(i.e. smartzone)? Are they really going to use, or need to use the dozens of talkgroups in place? 30+ million for a new Project 25 Motorola X2-TDMA system! I guess bragging rights these days cost that much.

Don't get me wrong it's very neat to see what they will come up with next. VoIP systems will and always will be vulnerable to the most basic attacks.Dos, DDos, PDos(say goodbye to your system if this happens) can cripple communications. Then what are you left with? Nothing more than a few analog failsoft(motorola) channels. That is if failsoft doesn't fail itself, which it has in the future. Can you tell I don't like motorola very well.

Take it or leave it, new more advanced systems are here to stay. Maybe weakness's should be worked on before moving on to the next best thing.

 

[hitechRadio]

You would be happy with Crystal controlled radio's I suppose. I guarantee the scanner u use to listen to P25 systems has firmware in it. And you might even use some software on your computer to program your scanner. See my point, The Firmware makes it work, the software just supports it. Same thing with these high tech systems you love to hate.

You lost me about your point on VoIP and (Motorola) failsoft. You are just making stuff up there..LOL

So if you don't like Motorola, can you tell us Y?
And what is your preferred radio of choice? Please don't say Yaesu.



Conventional or Trunking, Digital is here to stay, I am happy to say!!!

 

[JRayfield]

Don't forget the fact that just about every two-way radio made since the early to mid 80's has been firmware/microcontroller/microprocessor-based. Most "hardware-only" equipment disappeared from the late 70's into the early 80's.

I do have to completely agree with "Maybe weakness's should be worked on before moving on to the next best thing.". I've seen an awful lot of communications products that were sold before they were really 'ready'.

John Rayfield, Jr. CETma

You would be happy with Crystal controlled radio's I suppose. I guarantee the scanner u use to listen to P25 systems has firmware in it. And you might even use some software on your computer to program your scanner. See my point, The Firmware makes it work, the software just supports it. Same thing with these high tech systems you love to hate.

You lost me about your point on VoIP and (Motorola) failsoft. You are just making stuff up there..LOL

So if you don't like Motorola, can you tell us Y?
And what is your preferred radio of choice? Please don't say Yaesu.



Conventional or Trunking, Digital is here to stay, I am happy to say!!!

 

[MattSR]

Heres the final paper as presented. www.nicta.com.au/pub?id=5076

Section 5.3 - There is a Known Plaintext Attack. One of the oldest and most well known crypto attacks. Once you have a KPA, or a 'crib' as cryptanalysts call it, then your encryption scheme is no stronger than the algorithm its based on. DES can be brute forced in under a day - this is a fact and while expensive, is computationally straightforward.

It details how the attacks work. If you don't believe its possible after reading that paper, then theres not much further that can be done to help you. The information in there is based on scientifically proven fact. Counter arguments that consist of "Oh, no one can fly to the moon, so its obvious you didn't break DES-OFB" don't hold any weight..

Regards
Matt

 

[OCO]

Matt:
It would be interesting to hear of any responses to this presentation, especially if APCO ever steps up...I've been watching to see the same from the presentations of the University of Pennsylvania group. The only response I saw was from EF Johnson..

 

[molay]

I wonder if Mototrbo enhanced privacy , an unkown cipher AFK, is really just RC4.

Given mototrbo enhanced privacy uses a 40bit , the same as ADP (or WEP64 minus its 24 bits IV)

 

[JRayfield]

These are the kind of 'intelligent' questions that I love to see people ask. That's how people learn, instead of just 'assuming' and the posting incorrect information (and then getting mad and arguing when they are 'corrected' with facts).

Anyway....

You are somewhat correct on both accounts.

MOTOTRBO Enhanced Privacy is somewhat based on ARC4 (Alleged RC4). However, it has been 'modified' and the actual algorithms are kept 'secret' by Motorola. Not even licensed developers and partners, in their MOTOTRBO Application Developer's Program, have access to these algorithms.

However, the MOTOTRBO system planner gives a hint at what they did - rather than just using a 'static' key, changes are made in the 'key' information, for each voice and/or data packet that is sent over-the-air. The result is an encryption system that is kind of a 'digital version' of analog rolling code scrambling. This makes it very secure (obvsiouly, not as secure as AES encryption, but quite good).

John Rayfield, Jr. CETma

I wonder if Mototrbo enhanced privacy , an unkown cipher AFK, is really just RC4.

Given mototrbo enhanced privacy uses a 40bit , the same as ADP (or WEP64 minus its 24 bits IV)

 

[MattSR]

I wonder if Mototrbo enhanced privacy , an unkown cipher AFK, is really just RC4.

Given mototrbo enhanced privacy uses a 40bit , the same as ADP (or WEP64 minus its 24 bits IV)

Correct - its RC4.

 

[MattSR]

Nah - the changes are very simple - its the same deal as with P25 + ADP. Details are in the paper listed above.

 

[molay]

after reading the paper I wonder if is public the source code of the exhaustive key search attack on 40bit ADP?

Another question. How do you difference silence codewords of those that contain voice after encryption? leght? or simply is assumed that the first moments of communication must be silent?

 

[MattSR]

The source code isn't public as there is another paper in the works that uses a novel approach to break RC4 in much less time.

Like any other computer, the IMBE vocoder is a deterministic device. On our XTS5000s and XTS3000s It ALWAYS outputs silence at the start and end of a transmission, regardless of what audio is present at the radios input. For example - if I yell into the radio, and then key it up, the first 4 codewords and the last two codewords are still silence. This is repeatable and tested across many systems and radios - see the screenshots below.

 

[MattSR]

Also please read this Motorola Patent (Thats where the idea for the crack originated from about 18 months ago) - the silence codewords are deliberately inserted into the transmission after the transmitter key is released.. its almost like it was done on purpose...

Selective transmission of encoded ... - Google Patents

 

[MattSR]

And while I'm in a screenshotty mood, heres the receiver in action:-

http://mattrobert.com/Wireless/Screenshot.png

 

[SKEYGEN]

You talk about "P25 encryption," but exactly WHAT encryption format are you referring to? AES is Type I certified for protection of Top-Secret information. Some TS information is just tactically important (that a large air-strike is about to take place), but a lot has a shelf-life of many years, meaning the AES cipher key being broken even a year after the AES-protected transmission reasonably could be expected to cause exceptionally grave damage to US national security despite the secured communication not being decoded in REAL TIME by an unauthorized party. Thus, P-25 AES's goal is far from simply protecting the info from being decoded by an enemy in real or even near-real time.

Sorry, but you've mangled this completely. Type 1 applies to products, not algorithms. Among other criteria, Type 1 products implement classified Suite A algorithms. AES is a Suite B algorithm.

Certain evaluated and approved products, using TOP SECRET key material provided by NSA can be used to transfer TOP SECRET information, and the overall system incorporating these cryptographic devices is itself subject to TS accreditation. You can't just go to Best Buy and start using something that says it does AES to pass TOP SECRET information, the product must have been evaluated, you need to be using appropriate key material and it must be used as part of an accredited system. Suite B can't be used for all TOP SECRET information; information handled as part of certain other control systems at this classification requires the use a product implementing a classified Suite A algorithm.

 

[MattSR]

Isn't it amazing now that the proof is out there, all the nay-sayers and disbelievers have miraculously gone silent

 

[rdale]

Proof will be when realtime decrypted audio is posted, which has yet to happen...