E-tags cannot be cloned? Really?

daveza

Honorary Master
Joined
Apr 5, 2004
Messages
49,730
Reaction score
21,468
Location
Durbanville, Cape Town
http://www.itweb.co.za/index.php?op...53323&A=RFI&S=The+Spike&O=E&[email protected]

In recent media statements on the benefits of purchasing an e-tag to pay Gauteng's controversial tolls, the South African National Road Agency Limited (Sanral) claims that unlike vehicle licence plates, which can be cloned to fool photographic tolling systems, electronic RFID tags cannot be cloned.

E-tag egotism

It is the kind of over-confidence that leads to grave security disasters.

This is an astonishing claim. It is the kind of over-confidence that leads to grave security disasters.

It is not trivial to clone a radio frequency identification (RFID) device, but it certainly can be done. Simple breaches read the signal from an RFID tag and simply copy it. This remains vulnerable to systems that implement sufficient security to recognise and verify the unique tag ID, but the benefit is that the hack is easy to perform. More sophisticated hacks clone the entire device, including its tag ID. These are almost impossible for a reading system like the e-toll gantries to identify and reject.

Although the RFID industry's Association for Automatic Identification and Mobility claims that full tag cloning requires a bulky reader that can easily be spotted by security staff, vehicles travelling past e-toll gantries are not inspected for such devices. Moreover, the form factor argument seems to have been overcome in this example of an open source passive RFID tag cloner for access control cards.

It's not like RFID security issues are anything new. Various systems have, over the years, proved vulnerable to hacking, as this article from Wired Magazine in 2006 illustrates. They include tags designed to pay for services such as fuel.

When Mythbusters, a show on the Discovery channel, proposed to test RFID for security, trackability, and reliability, matters got rather heavy, with big-gun lawyers from the financial industry leaning on the producers not to air the show. Here's Mythbuster Adam Savage describing that incident. If they were so confident that RFID tags could not be cloned, hacked, spoofed or otherwise subverted, one expects that they'd be proud to have renowned tech geeks like the Mythbusters have a crack at it.

The fears of the credit card industry are well-placed. Stories of cloned credit cards surface regularly, and staying a step ahead of the criminals who exploit vulnerabilities is a full-time occupation for even the most sophisticated experts in the financial industry.

It is true that the Sanral e-tags have several layers of security, including being matched to specific vehicle number plates and types. However, a group of activists known as the Opposition to Urban Tolling Alliance is already in court arguing that the system is so complex it is inconceivable that it could be administered effectively. Millions of registered vehicles will whizz past tolling gantries, paying as much as R10 million per day in tolls. The system would issue 840 000 invoices and 30 000 summonses per month, which would pose “insurmountable logistical problems”, the group claimed in interviews with the Saturday Star Motoring.

Even if these checks could be made consistently, however, and disputes settled efficiently, it is not inconceivable that more sophisticated fraud could succeed, as it frequently does against financial institutions.

In a world where millions of credit cards are subject to cloning, despite the extensive experience and high sophistication of the industry that produces them, for Sanral to claim e-tags will be immune is stupid, misleading, or both. Either way, it is the sort of pride that comes before a fall. A spectacular fall, if Johannesburg's billing crisis is any indicator.

I guess the thing to do is for someone to actually clone an e-tag which would at least put a few more nails in this toll road coffin.
 
So, I just clone your etag and your numberplate at the same time.... hell its not that difficult to walk past a vehicle, clone the tag and write down the number plate... go off and get a set of plates made...

Anyone with some technical knowledge would have seen this rather massively gaping HOLE in the etag saga from day 1... but I'm guessing it was buried, because it didn't suit the gravytrain to hear it.
 
So, I just clone your etag and your numberplate at the same time.... hell its not that difficult to walk past a vehicle, clone the tag and write down the number plate... go off and get a set of plates made...

Anyone with some technical knowledge would have seen this rather massively gaping HOLE in the etag saga from day 1... but I'm guessing it was buried, because it didn't suit the gravytrain to hear it.

While it will not completely block it, it does make it harder - not everyone will have an e-tag... but yes, you do have a valid point.
 
Anyone offering such a cloning service yet? I'd spend money to screw with their system.
 
Anyone offering such a cloning service yet? I'd spend money to screw with their system.

1. RFID reader: check the info it pulls.
2. Write said info to a passive tag (e-tags are passive, not active, afaik, so they get power from the reader)
3. Test by driving past tollbooth :P
 
What I'm hoping for is this:

A combined check, as mentioned in the article: an e-tag combined with a numberplate. In other words, for a transaction to complete successfully, an e-tag should only be billed for that user if the registration number from the photograph matches. If not, it should go int a separate pile for other checks (ie e-tag without numberplate, and then checked against vehicle, etc). This would make it necessary to clone both the plates and the tag for the clone to work.

For anyone with the necessary know-how and RFID reader, it's trivial to clone an e-tag. It's as simple as that. The "equipment" isn't bulky either. It can be as big (or small, rather) as a cellphone.

I would go further for additional checks like vehicle colour, shape, make, model - also some logic around simultaneous reads that would be physically impossible along with improbable reads (like a car going from Pretoria to JB, then suddenly on the same day from Centurion to JB after already logging a normal trip). At least then you would need to clone the tag and numberplate, you would need to use it on the same type of car in the same colour and you would need to manage your routes carefully to use the clone. Cost and complexity of the illegal route starts getting too high then.
 
I would go further for additional checks like vehicle colour, shape, make, model - also some logic around simultaneous reads that would be physically impossible along with improbable reads (like a car going from Pretoria to JB, then suddenly on the same day from Centurion to JB after already logging a normal trip). At least then you would need to clone the tag and numberplate, you would need to use it on the same type of car in the same colour and you would need to manage your routes carefully to use the clone. Cost and complexity of the illegal route starts getting too high then.

Yes. However determining colour at night is not possible. (Only possible with colour cameras and these only work during the day. Numberplates are read using infra-red and/or black+white).
Determining make/model is still in a development phase with number recognition software.

Unfortunately attempting to monitor improbably routes is not simplistic - there would nee to be a database that gives you such improbabilities and the load that this would put on the system would be immense, if not unsustainable.
 
Last edited:
What I'm hoping for is this:
For anyone with the necessary know-how and RFID reader, it's trivial to clone an e-tag. It's as simple as that. The "equipment" isn't bulky either. It can be as big (or small, rather) as a cellphone.

Not necessarily. The e-tag system could be using a PRNG so that the information sent each time differs. I'd be more than happy to try and see what it's made of but I don't know anyone with an e-tag in Cape Town :p
 
Not necessarily. The e-tag system could be using a PRNG so that the information sent each time differs. I'd be more than happy to try and see what it's made of but I don't know anyone with an e-tag in Cape Town :p

Okay so I spoke to someone that's an expert in the field...
In short what he said:
a) Technically it's possible
b) It will be hit and miss with the protocols and the technical details.
c) At the end of the day it will be a big b****h to clone.
 
Okay so I spoke to someone that's an expert in the field...
In short what he said:
a) Technically it's possible
b) It will be hit and miss with the protocols and the technical details.
c) At the end of the day it will be a big b****h to clone.

If they use a PRNG, it will be almost impossible (save for brute-forcing) unless the seed can be obtained or they have the IV vulnerability in RC4. Reverse engineering the protocol is the least of your worries then.

Or, all they have to do is just hash a date-time stamp + some secret keyword and use it as a checksum. Good luck replicating that without the keyword.

Fact is, it could be the simplest thing of unencrypted license data or it could be what I said above. Only one way to know and that's through having one of these e-tags to experiment with.
 
Last edited:
Yes. However determining colour at night is not possible. (Only possible with colour cameras and these only work during the day. Numberplates are read using infra-red and/or black+white).
Determining make/model is still in a development phase with number recognition software.

Unfortunately attempting to monitor improbably routes is not simplistic - there would nee to be a database that gives you such improbabilities and the load that this would put on the system would be immense, if not unsustainable.
I was thinking using facial recognition type software to make out car shape as a minimum, but fairly good starting effort. Therefore SUV profile vs sedan vs hatch vs minibus and so forth. Colour at night would be an issue, but then again there are also supposed to be decent street lights...

From a technical point of view the etag thing is great as it has so many problems to try solve = money pit.

It just needs one to make the system vulnerable.

I am sure they can ask Postbank for a few tips
 
Top
Sign up to the MyBroadband newsletter
X