Possible Serious Security Flaw In ATMs

sfjoe writes "According to a story at MSNBC.com, researchers at Algorithmic Research (ARX) have shown it may be possible for 'someone with access to the ATM network to attack the special computers that transmit bank account numbers and PIN codes, called hardware security modules'. Using these methods, an attacker could trick the security modules into exposing a PIN. It has long been considered impossible to access PINs as they are traveling through the ATM network without the encryption key used by the card-issuing bank. If PINs can be compromised, the almost 8 billion transactions per year they handle may be in danger. Not to mention all the transaction at retail stores."

Comments

Intercepting Transmission

[DigitalRaptor]

I saw a news report the other day of a guy that hooked his a device (it may have been an iPod) to the back of an ATM where the phone line comes out, and intercepted the signal transmitting the information.

He was able to get credit card numbers, pins, and all of the other information transmitted, and stole a lot of money before being caught. And he wasn't caught by bank security or software, he was caught because a clerk was paying attention, IIRC.

Re:Intercepting Transmission

[DigitalRaptor]

This one.

Also covered here.

And here.

If there was crypto used, it absolutely sucked.

If all you need is a modem line tap or an illegal program to crack ATM's, there isn't much security is there?

I don't think there is crypto. I think the information is sent across the phone lines as plain text. The purpose of the modem line tap or illegal program is to convert the signal going over the line (the same signal you hear when you pick up the phone during a fax or internet connection) to text. From there, no mention is made of encryption.

See this page. "The Modem Line Tap, MLT2400A is a modem protocol analyzer that translates telephone data communications into standard ASCII characters for display on a PC screen."

If the data was properly encrypted before it was sent, the hackers wouldn't have been able to use the data. If there was crypto, it was token crypto at best. Just enough to tell their share holders it was encrypted.

Holding All the Cards

[Doc Ruby]

Every bank I know of with back-end offices here in NYC requires everyone passing through their building doors to use onetime password cards (usually RSA keycards) for access. Yet those banks all make us run around broadcasting our PINs to whichever fly-by-night ATM dispenses $100 latenight when we're drunk.

The cost of chipcards that generate onetime passwords, to protect from replay attacks, is minimal. Especially compared with fraud and theft. What's taking them so long?

So just use it as a credit card?

[letsgolightning]

I realize this topic is mostly meant for using a card at an atm to take out cash and the like, but whenever I use my debit card to actually buy something, I make sure to use it as credit, even though most stores' touch-and-swipe pads love to default to a keypad to enter a pin. I just hit 'cancel' then 'credit' and sign the screen. No pin gets transferred, so I don't have to worry about anyone stealing it. Usually, they ask for an id because my signature is so awful (added security for me). I get points for my purchases, which I may be able to redeem within the next decade. And best of all, if anyone does decide to defraud me this way, Visa and my bank will give me the stolen funds back (my bank covers the $50 or so 'deductible' that Visa normally wants). To quote Micheal Scott, it's a win-win-win. I'm safer, my money's safer, and Sam Walton gets less profits because he now has to pay Visa processing fees.

Re:Poink-Poink-Poink-Poink

[TubeSteak]

The thing is, anyone can purchase & setup an ATM.

There's almost no State/Federal regulation (that I'm aware of).

Re:The reality of this is...

[ZzzzSleep]

Quoth Chosen Reject

But if tellers ever get to the point that store clerks do (and I suspect many have) then any old schmoe will be able to take money out of your account. I can't tell you how many times I've had cashiers ring up a sale without ever even looking at either my ID or my signature on the back of the credit card. I've had times where I offered and was refused, as if they didn't want to have anything to do with security checks of any variety as that might bring upon them responsibility or something. I'm not talking about small purchases here either.

Have a look at Zug's credit card prank http://www.zug.com/pranks/credit/ and be afraid for the security of your funds...

Re:Poink-Poink-Poink-Poink

[statusbar]

And that is why organized crime has their own ATM division:

http://www.beyondrobson.com/tech/2006/10/avoid_ban k_fraud_i_didnt/

Therefore, not only is the ATM network insecure, it always has been for other reasons.

--jeffk++

What a coincidence!

[Mark_in_Brazil]

I'm sure it's just a coincidence that Algorithmic Research (ARX) is a vendor of security solutions, including HSMs , and that ARX has been losing market share in that space for years and has a tiny market share (nCipher dominates the HSM market worldwide, Safenet, through acquisitions, has the next-largest market share, and then you start getting to competitors with very small market shares). I'm sure the researchers at ARX had no idea that almost all banks in the world use HSMs made by competitors of ARX and just wrote this paper to expose a very real security flaw, one that something tells me ARX HSMs don't allow...
FWIW, ARX was actually something of a leader and had some cool ideas... several years ago. I'm not sure whether it was because of financial trouble, incompetent management, neither, or both, but they were lapped by players like nCipher, Luna (now part of SafeNet), Utimaco, even Thales, which focuses on serving the credit card transaction market but doesn't have things like Diffie-Hellman key exchange because VISA and Mastercard don't require them, and yes, even the old low-cost option, Eracom (bought by Safenet in order to do away with a pesky competitor).

Really Unlikely...

[fixer007]

I work at a 'switch' that the article describes. It would be REALLY hard to do what they are describing, even having inside access. Not to say it couldn't be done, but the person doing it would have to have some serious clearance to get access to the HSM and the system it is on. If they do have that kind of access, it is pretty unlikely that they have the technical know-how to go about doing what the article describes.
Usually the people that have the technical know-how don't have userid's or passwords to the production system, never mind the HSM.

I would be much more worried about someone with a hidden camera getting your PIN in a gas station than this. Alot cheaper and easier to pull off.

Who knew the system was this broke?

[Kadin2048]

My thoughts exactly.

There must be some reason (I hope) but the security model that they're describing in TFA seems horribly flawed. It depends purely on the security of some black-box hardware modules embedded at different points in the system.

Basically, what they're saying is that there's no end-to-end encryption of your "PIN block" (PIN+Account number, don't ask me why they're transmitted together instead of separately with some random transaction identifier). Instead, the ATM encrypts it for the next machine in the network, where it's decrypted and re-encrypted inside an (assumedly secure) hardware module. Then it's passed to the next link in the chain, ad infinium.

This wouldn't be bad, if the ATM first encrypted the PIN block using the public key of the eventual destination bank -- after all, the intermediate machines have no reason to actually know your information, they're just shuffling bits. However, to just use this transmission-level wrapper without actually encrypting the data seems horrifically stupid. It's nothing but 'security through inconvenience.' (It's not exactly even obscurity, since people seem to know how the system works, they just make it inconvenient to intercept the information by making the places where it's unencrypted relatively small.) From a crypto perspective, it's a broken system.

Re:The reality of this is...

[Majik Sheff]

It's mostly historical. The ATM network was put in place when DES was considered an effective encryption algorithm. DES was nice because it could be easily implemented with a handful of logic gates and it was a public algorithm. It was also reasonably secure, when people couldn't buy a few gigaflops at CompUSA. Now the ATM network is just finishing up an end-to-end upgrade to 3-DES and I doubt the world's banks will be in a hurry to obsolete their networks again any time soon. Public key encryption is the right choice, unfortunately it's not the choice that our banking overlords have made.