Hacker News Reader

Comment by wschroed | Hacker News Reader

How do chips solve the problem? From my understanding, a man-in-the-middle scanner can gather enough information about query-responses to simulate the chip. This was one of the big problems identified with RFID chips embedded in passports because all a criminal would need to do is brush by other travelers with his skimmer.

View on HN

I've seen two different kinds of reports:

* Some banks cards were vulnerable due to faulty crypto. The banks phased those cards out.

* Attacks based on a malicious PIN pad logging the PIN code, then feigning a chip error and telling the user to fall back to the magstrip, thus turning to traditional skimming.

I haven't read anything that attacks the chip itself on current cards. Do you have any links?

edit; Just found http://en.wikipedia.org/wiki/EMV#Vulnerabilities

edit2: Wikipedia TL;DR: There are two currently-relevant attacks:

* One lets attackers trick a terminal into initiating a PINless transaction in order to use a stolen card. This information is sent to the issues as part of the authentication, so a bank could deny all PINless chip charges if they wished (I'm not sure what cases this legitimately used in?), plus there's a clear trail that the cardholder isn't liable.

* The latest attack tricks the card into downgrading to an older, plaintext method of transferring the PIN from the terminal to the card, allowing the PIN to be skimmed. I'm not sure how this is useful in recreating the card to steal money.

There's some evidence that the first attack was used in the wild, but the banks deleted the logs showing whether a PINless transaction took place so the customers were found liable for the charges.

To make matters worse, after C&P rolled out some banks would just flat out refuse to reimburse cardholders for fraudulent charges. They claimed the system was bulletproof and if fraud did happen, then it was the cardholder's fault: http://en.wikipedia.org/wiki/Chip_and_PIN#Banks.27_liability

Unfortunately it's regular practice to say that things are bulletproof and airtight :/

The chips are sophisticated and powerful enough to setup an SSL/TLS session to the bank if needed. Just like with SSL/TLS, if it's designed correctly, intercepting the traffic between the card and POS/ATM is useless.

I haven't been able to find very much on this in a few evenings that I searched for info on it, but from what I've been able to find I'm quite sure that it's possible to do. The chips give a boolean response as to whether the PIN number is correct and lock themselves after 3 attempts. Combined with the knowledge that 6 pins (=6 parallel bits) are used for 4 digit PIN numbers (log(10^4)/log(2)=14bits), you can deduce that it must have at least persistent storage and computational capabilities. A complete Von Neumann machine. This makes it feasible to implement algorithms like RSA and AES (asymmetric and symmetric encryption) on the chips. The POS/ATM then provides power and a connection to the desired bank, and all should be fine.

I don't really think they are that good to connect to the bank directly, but the idea that the chip is capable of this kind of crypto makes me feel better than with magnetic strips.

They are "sort of" connecting to the bank - the idea is that the card issues an authorisation token w. the amount and other info, signs it with the private key, and then the merchant sends it to the bank which may approve or deny the transaction.

The chip is authenticated with a public/private key challenge/response. The private key is never sent to the ATM; the chip actually runs a very small program/system that can generate the correct response to the challenge using its private key. The chip is powered by the contact with the ATM.

Physically how would you perform a man-in-the-middle? I think you're confusing the chip-and-pin chip (EMV) with an RFID chip.

EMV is a contact system that requires physical metal pins to touch multiple contacts on the top of the chip itself.

No, it can't. The chip works with a challenge/response system (and this only works after the pin is input to the chip)

What happens is that they clone the magnetic strip, so sometimes that and pin is enough to produce a card that does withdraws.

Reply on news.ycombinator.com