Researchers Find ‘Backdoor’ in TETRA Encrypted Police and Military Radios

[dimab]

Yikes. Hopefully this post stays up. It's in the general tech news. Not intended to start a decryption debate.

Researchers Find ‘Backdoor’ in Encrypted Police and Military Radios

The TETRA standard is used in radios worldwide. Security researchers have found multiple vulnerabilities in the underlying cryptography and its implementation, including issues that allow for the decryption of traffic.

www.vice.com

 

[ecps92]

Yikes. Hopefully this post stays up. It's in the general tech news. Not intended to start a decryption debate.

Researchers Find ‘Backdoor’ in Encrypted Police and Military Radios

The TETRA standard is used in radios worldwide. Security researchers have found multiple vulnerabilities in the underlying cryptography and its implementation, including issues that allow for the decryption of traffic.

www.vice.com

Glad TETRA has not taken off in the good ole US of A, then.

 

[Pr999]

Glad TETRA has not taken off in the good ole US of A, then.

Guess what, it has!

 

[GTR8000]

Guess what, it has!

It has? There are around a dozen TETRA systems in the U.S., most of them small systems at airports or refineries. The two largest are the NJ Transit and NYC Transit Authority systems. I would hardly call that "taking off" compared to the hundreds and hundreds of P25, DMR, NXDN, etc. system types in the country.

 

[Pr999]

It has? There are around a dozen TETRA systems in the U.S., most of them small systems at airports or refineries. The two largest are the NJ Transit and NYC Transit Authority systems. I would hardly call that "taking off" compared to the hundreds and hundreds of P25, DMR, NXDN, etc. system types in the country.

well i'm saying 'technically' it has.

 

[KevinC]

TETRA is a spectrum hog. The signal is just as wide as iDEN so it's nearly impossible for anyone to use an adjacent channel. And that's a 25 kHz channel, not 12.5.

 

[jtwalker]

well i'm saying 'technically' it has.

A few agencies/municipalities with more money than they know what to do with, doesn’t really constitute “taking off” imho.

 

[BinaryMode]

Very interesting. Question: is P25's use of AES256 a closed, proprietary crypto? I mean its implementation?

 

[mmckenna]

Very interesting. Question: is P25's use of AES256 a closed, proprietary crypto? I mean its implementation?

AES256 is not a closed/proprietary crypto. It's supported by all the big LMR radio manufacturers. It is import restricted.

The keys, on the other hand…..

 

[BinaryMode]

LMR? Land Mobile Radio?

 

[mmckenna]

LMR? Land Mobile Radio?

Correct.

 

[GTR8000]

well i'm saying 'technically' it has.

Well, you're wrong. A couple of systems is far from "taking off", technically or anecdotally.

 

[Ubbe]

TETRA is a spectrum hog. The signal is just as wide as iDEN so it's nearly impossible for anyone to use an adjacent channel. And that's a 25 kHz channel, not 12.5.

It's a spectrum hog but Tetra are often assigned to reserved frequency ranges where no other communication are allowed. It has 4 time slots in a 25KHz wide channel so you can see it as 6.25KHz per slot. It also has a fantastic audio quality, in most cases you can't tell is digital.

The title are misleading as it says they have found a backdoor in police and military radios. The weak encryption are found in TEA1, the lowest level of encryption that are more or less just a scramble function. Police and military use TEA2 that haven't been found to have any weakness in its encryption. I haven't seen anyone actually using TEA1. The Tetra systems are either in the clear unencrypted or use TEA2 as TEA1 are something that's only useful for blocking monitoring programs like a Tetra plugin for #SDR as it doesn't cost anything to use TEA1 but are instead not secure enough to be used for critical users.

/Ubbe

 

[kayn1n32008]

Very interesting. Question: is P25's use of AES256 a closed, proprietary crypto? I mean its implementation?

AES256 is NOT closed or proprietory. In LMR it is used and supported by a majority of manufacturers.

Just the opposite, it's open, and been repeatedly tested for weakness/exploits.

 

[maus92]

TETRA is a spectrum hog. The signal is just as wide as iDEN so it's nearly impossible for anyone to use an adjacent channel. And that's a 25 kHz channel, not 12.5.

TETRA also does not support simulcasting, adding to its spectrum inefficiencies.

 

[R0am3r]

A Wired article provides a few more details about the TETRA TEA1 hack. "All four TETRA encryption algorithms use 80-bit keys, which, even more than two decades after their release, still provides sufficient security to prevent someone from cracking them, the researchers say. But TEA1 has a feature that reduces its key to just 32 bits—less than half the key’s length. The researchers were able to crack it in less than a minute using a standard laptop and just four ciphertexts."

Code Kept Secret for Years Reveals Its Flaw—a Backdoor

A secret encryption cipher baked into radio systems used by critical infrastructure workers, police, and others around the world is finally seeing sunlight. Researchers say it isn’t pretty.

www.wired.com

 

[RayAir]

TEA1 and 4 are approved for export.
90's proprietary encryption.
Not surprised at all.

Reminds me of Motorola DVP. Has a very large key variable on paper but turns out it is linear and after analysis there were only 255 (approximately) possible keys.

I'd be highly suspect of any proprietary encryption:
Motorola DVI-XL, DVP-XL
GE/Macom VGE

I would love to see an analysis on DVP-XL and VGE.

Thankfully today open encryption standards are used (AES).

 

[Ubbe]

TETRA also does not support simulcasting, adding to its spectrum inefficiencies.

I'm not sure that simulcast are that more spectrum efficient over a large area. You will need to have a high capacity for the city area to handle all its users. And then you'll need to build just as high channel capacity in the suburbs where there are perhaps 5 times fewer users. In a non-simulcast system you can use a high capacity number of channels for the city but in the suburbs you only need 1/5 of the number of channels to not get a busy. And at the other side of the city in the suburbs you can reuse the frequencies from the other suburb area across town without interference issues. You free up 4/5 of the frequencies at the outer side of the suburbs to be used for other purposes. In simulcast you block the frequencies from being used by others in a very large area.

It probably depends a lot of the terrain and how directional antennas are being used to favor a simulcast solution. I hope they use unbiased third party evaluaters before deciding for a system design as the radio companies will suggest a system that will sell as much hardware as possible. Simulcast systems are a real money pit that needs good finances.

The main reason for simulcast must be that radios do not have to roam between sites, it's seamless between sites and the control channel will then have a very light load when stripped of roaming info. For a Tetra system the control channel data can be so excessive that it needs users to roam to secondary data channel slots if they transmit a lot of GPS positions and other additional info and at the same time switch sites at a high frequent rate when moving between high rise buildings.

But I can see that in simulcast the control channel of a site will be interfered from users not local, as they use the same frequency, and will load the control channel and asks for repeats of data from local users that gets interfered. So perhaps the interference level for site receivers might be higher in simulcast that reduce the max possible load level.

/Ubbe

 

[radionx]

"Thankfully today open encryption standards are used (AES)."

Correct implementation is relevant. The whole chain has to fit.

And no, I do not trust anybody. Encryption won't save a failing society under hybrid warfare.

Would be very sad if the chinese and russians knew about your backdoors...

 

[mikenet]

TEA1 and 4 are approved for export.
90's proprietary encryption.
Not surprised at all.

Reminds me of Motorola DVP. Has a very large key variable on paper but turns out it is linear and after analysis there were only 255 (approximately) possible keys.

I'd be highly suspect of any proprietary encryption:
Motorola DVI-XL, DVP-XL
GE/Macom VGE

I would love to see an analysis on DVP-XL and VGE.

Thankfully today open encryption standards are used (AES).

If anyone is interested, here‘s that paper on DVP (non-XL):

https://www.cryptomuseum.com/crypto/motorola/saber/files/vulcan_201409.pdf