I have also designed security protocols to keep the data and keys from the bad guys, including things like Key Injection Facility, loading and exchange protocols, etc.
These things (terminals, cards, HSMs) can be extremely safe. Not only there are protocols available that in my knowledge have never been breached but also the devices themselves are built in secure manner to prevent any physical attacks. The entire infrastructure can be made so that no single employee can get hold of any secret material.
The truth is, that there is a range of solutions and the secure ones cost. It costs to set them up and use correctly. It costs to produce chip card that can dynamically sign and encrypt the data. It is up to the issuer (bank) how much they pay for the security of their card and typically banks are not too eager to overpay. Rather, they take calculated risk.
Another problem is that North America is particularly behind in security standards. I remember, while working on this application (10 years ago) there was much more leniency for US users (issuers, acquirers) than the rest of the world.
They aren't 'cracking' Key Injection Facility or messing with protocols. They just add a simple card reader? The reason they hit gas machines is the volume and most don't take chips (US).
https://www.amazon.com/dp/B01N4L19CR/ref=asc_df_B01N4L19CR51...
For the fraudster to receive any funds from my account they would have to go with that data somehow to my issuing bank and have their transaction request accepted.
My bank will never do this because my card is a secure one. They expect it to generate a signed document listing the details of my transaction along with the name (ID) of the merchant and terminal performing the transaction. The message also certifies that the PIN I entered was successfully verified by my card. The message can only be created by the card using my PIN. This requires the card to be stolen along with the PIN.
The only money I can loose in case my card is stolen without the PIN is up to a floor limit for offline transactions. It's peanuts but it allows for quick transactions for example from vending machines.
The gist of the story is to move to a bank that cares for security.
Forgive my ignorance, but aren't most sellers in America willing to process transactions (even large ones) without PIN entry?
I realize pin-and-chip is a largely secure scheme, but the chip rollout in the US hasn't actually implemented a two-factor system.
http://www.emv-connection.com/best-practices/what-is-the-fra...
https://usa.visa.com/visa-everywhere/security/emv-at-the-pum...
While I don't doubt they're already out there, I don't recall ever seeing a gas pump in the US with a chip reader.
Also, it's at leas worth noting that technology for skimming chip-based transactions already exists in the wild:
https://krebsonsecurity.com/2017/01/atm-shimmers-target-chip...
The "shimmers" take advantage of poor implementations of the EMV standard, but the chances strike me as fairly high that there's going to be a lot of poor implementations of EMV out there in the wild. (Not only is the less secure "chip and signature" approach common in the US, I've made more than one purchase with what could be described as "chip and nothing": put my card in a chip reader and make a purchase without being asked for either a PIN or a signature.)
Buy stolen CCs without pin in large volume
Set up online merchant account (eBay is common), use software to scrape and copy listings for popular products that cost less than the floor amount but more than your costs. When someone orders a product and the payment clears automatically put in an order for the product on a different website (preferably one that allows a guest checkout). Automate more and more parts of the process as time goes on (e.g. using a different card when one is declined)
Keep the transaction volume low and you're basically printing money at the rate of the difference between card info cost and item cost.
This is a pretty common tactic to monetize large volumes of CC#s.
It doesn't help that most merchants don't cross reference billing address with the bank and some don't require the security code.
Two years ago at defcon, someone presented on advanced skimmers that could read chips too. He claimed they need a network connection to work, as there's only a two minute window to reuse the same token.
This was noted in the article. There appears to be availability on the board for an extra connection.
Not quite. The one who gets your card info probably won't use it. They will more likely sell it. If the person who buys it can't use it, whoops, too bad.
And credit-cards are a testament to that.
1) transactions with card present(chip+pin or signature)
2) transactions with card not present(someone types in your card number into the terminal)
3) internet transactions(card number + CVV from the back).
I can set any of them to zero, effectively blocking the transaction type completely. So yes, while you can read the card number and exp date from the chip, you can't get the CVV number from the back of the card, which means it's only useful for the second type of transactions - and I don't see any reason to ever change the limit on that to anything other than zero - so any information gained from just skimming the chip/magstripe is effectively useless without the CVV written on the back.
The better solution is to have a bank with good reputation that expends effort in making your transactions safer.
My bank:
- has their cards well issued (and I know this since I have worked as credit card terminal developer for many years and I know every detail of how cards are personalized and designed actual security systems),
- sends me a code to verify my internet transactions (remember not to use your phone to do internet transactions!),
- processes chargebacks without fuss unless there is a reason to suspect cardholder is trying to defraud the merchant.
Do you have time to explain?
I'm aware of the problems with using sms as 2FA but if I understand you correctly you mean something else?
(Around here most banks demand we use BankID which is typically downloaded to the sim card in the phone.)
If you were to enter your credit card details on the same device you use to receive codes (most likely your only mobile phone), the attacker having some kind of malware code could first steal your card information and then use your phone to receive the codes to complete the transaction.
This requires infecting just one device, so basically as a fraudster you create a malware and wait for people to have their phones infected. Then you defraud those that use phones for credit card transactions and either don't need separate codes to complete the transaction or use the phone for this.
It is much more difficult to get two devices infected that are used by the same user. This only typically happens in case of targeted attacks and is rarely seen.
My colleague at one of the companies I worked for lost the money he saved to buy a flat this way. He got his phone infected with malware and then over few days he got all his money sucked out of his account in a series of increasing transactions.
Now I keep a low-limit branded gas card that can only be used at gas stations. Go ahead and skim me. All they can do is charge $30 worth of Munchos and beef jerky.
The thinking is if either one of our cards get compromised we will can just replace that card, the third card will remain safe while locked up. That is unless the company or payment processor gets breached which is a larger issue in itself.
For recurring payments I make a ShopSafe card per merchant. If one merchant has a breach all I should have to do is make a new ShopSafe card for them. All of the others should be unaffected.
If the underlying card is compromised and replaced the ShopSafe cards should continue to work as long as the underlying account is not closed.
This has worked out quite well. There are only three annoyances.
1. The interface is annoying. It is a small Flash window.
2. The expiration date can only be up to one year out.
3. Some merchants, like Amazon, use a variety of merchant names depending on what you are buying. For example, the shoelaces I bought recently from Amazon show up on my credit card statement as "AMAZON MKTPLACE PMTS AMZN.COM/BILLWA". The Kindle book I bought recently comes from "Amazon Services-Kindle 866-321-8851 WA". My subscription to the Kindle edition of "Analog Science Fiction and Fact" comes from "KINDLE-AnalogScien 866-216-1072 WA". (A few months ago, Analog came from "AMAZON DIGITAL SVCS 866-2866-216-1072 WA"). I also see in my Amazon purchase history things from "Amazon.com AMZN.COM/BILLWA".
Dealing with that would require five ShopSafe cards, and picking the right one for a given purchase.
There is a workaround I've heard for this with Amazon, but have not gotten around to trying. Buy Amazon gift cards with a ShopSafe card, and use them to pay for everything else.
[1] Bank of America. I believe that Citi has a similar feature.
We don't touch the debit card attached to our checking account unless there's no choice, however - which is usually just WinCo when we can't get what we want (or get what we want for a reasonable price) at Fred Meyer or Albertson's. All of our credit cards have $0 liability, and worst case if someone skims them and maxes our limits we have to temporarily switch to a backup card while the charges are reversed and a new card is mailed - someone draining my checking account leaves me with the inability to pay rent, electricity and gas and Wells Fargo is happy to drag their ass on reversing unauthorized debit card payments.
My cards have been compromised by Target, Home Depot, and other vendors. I never know exactly who, just get the notice from the bank that I'm getting a new card.
The side benefit for me is that these small transactions keep the cards active with regular on-time payments.
If one of them is compromised, I only have one or two recurring payments instead of having to change all of Netflix, Ting, Comcast, Virgin Mobile, Google Play, Apple, NY Times, WSJ, etc.
I've probably had 5 cards compromised in the past decade. The biggest hassle is changing recurring payments. I don't track them, I just do my best ten wait for the bills to fail. What's annoying is sometimes they keep successfully charging my account for 4-5 months before failing. How they're able to bill the old number I don't know.
But yeah, they just call you and send you a new card. No need to select which transactions are fraudulent typically -- or at best 2 mins on the phone. No big deal.
Sucks when you're overseas but google voice + headphones and then Apple Pay means your new card is active the minute it's issued, even if you won't get it in the mail for 2 weeks.
[1] https://usa.visa.com/dam/VCOM/download/merchants/visa-accoun...
[2] https://www.mastercard.com/us/wce/PDF/Billing%20Updater%20Br...
As a consumer using a debit card, you have lots of exposure. Bounced transactions, lots of paperwork and more bureaucratic pitfalls.
Credit cards with premium issuers are just a minor hassle. With Amex or Citi, you literally just need to click. With lousy issuers, you can be stuck dealing with a breach event for a long time.
Where do you think the cost of losses goes? To the consumer via higher prices...
By the way, do you have any reference on this? Everywhere I've searched, chip seems synonymous with EMV.
Every chip or contactless transaction is an EMV transaction. This is by virtue of EMV being the protocol that terminal uses to communicate with the card.
EMV allows for different types of transactions with different level of security.
For example, if the transaction can be done offline (gas pump unattended terminal situation) AND executing CVM rules (both terminal and issuer have their own rules) points to the possibilty of exchanging encrypted PIN, the PIN can be sent safely encrypted to the card for verification.
If the card or the terminal does not support encrypted verification then the PIN can be sent in the clear and vulnerable to skimming.
> Every chip or contactless transaction is an EMV transaction.
But the parent comment I was replying to says:
> Just because you are using the chip doesn't mean you are doing an EMV transaction.
I presume you're saying he's wrong in this? Or am I misunderstanding what you both mean?
I'm also confused by the PIN thing you mention. Why is there a PIN at all, and why is there an "encrypted PIN exchange"? We're talking about the US, right? The US is chip & signature, and there is no PIN involved. I thought it was supposed to be that the card generates a unique one-time authorization code (public-key signatures?) and the bank validates it? Where does a PIN come into play?
Card Verification
- no verification (just read magstripe or magstripe equivalent from chip or contactless card that does not have any more advanced verification mechanism)
- static verification - for chip, card sends static signature along with the data so that the terminal can verify electronic signature
- dynamic verification - for chip, card accepts challenge and generate response so that the terminal can verify the signature. This also makes it much more difficult to copy the card because it is not enough to copy the available data, you also need to copy the key that is embedded in the card.
Cardholder Verification
- no verification -- sometimes no cardholder verification will be performed (for example contactless under certain limit, unattended terminals without PIN capability or on a plane when there is special rules for airline terminals because people are typically verified separately)
- Signature -- this is where US is stuck it seems
- offline plaintext PIN -- cardholder enters pin, terminal sends the pin to the card, card responds if the pin is correct -- this is the source of most of the skimming problem
- offline encrypted PIN -- same as above but the pin is being encrypted with a key established securely with the card. This is safe but the cards cost more.
- online PIN -- the PIN is never exchanged with the card, it is encrypted and sent to the bank and bank decides whether it likes it or not.
Transaction verification
- floor limit -- sometimes the transaction can be agreed between the terminal and the card. Typically there is some information stored on the card and set of rules that decide that this is possible. The card may be decreasing a limit of funds available offline and when it hits the limit it will force you to perform full chip transaction.
- online verification -- the message goes online to the bank and bank decides.
>- offline encrypted PIN -- same as above but the pin is being encrypted with a key established securely with the card. This is safe but the cards cost more.
what's the difference between the two? can't the attacker put a physical keylogger on the pin pad?
The problem comes from physically stolen cards, if your card doesn't rely on cryptography to secure the request/response channel you can insert a shim between the reader and the card to fake acceptance of an arbitrary pin. This specific attack has already been demonstrated, and if my memory serves correctly it's already being used in the wild.
I'm in the US and the credit card I have through my credit union requires a PIN on chip transactions.
Its all a stopgap measure to chip and pin anyway. Once everyone is used to EMV and all the terminals have been replaced, we'll have a sunset on swipe and then move to chip and pin, which will be a minor upgrade. This will all take time. Lets remember the United States has a quarter of the world’s credit card transactions, but only 4.5% of its population. Its a very big cc market and change isn't going to come quickly.
After about 3 cycles, it gives up and accepts the magstripe (plus the pin)
Often it will just deny the transaction completely and the shop queue grows _looong_
Later, before the new card arrived, I tried using the old one again and it worked, so I just chalked the whole thing up to those two stores likely using the same payment processor that was having issues that day or something.
What do you do when the chip and PIN fails?
[1] https://www.cardfellow.com/quick-chip-slow-chip-card-transac...
Since all of the laws revolve around signing, we don't get PINs.
There is a second wrinkle to this for Europe: because there is no equivalent to these laws, in many cases Europeans are the ones completely on the hook for fraudulent transactions as the correct PIN number is seen as evidence the consumer having leaked it. So in cases of fraud the U.S. chip-and-signiture winds up protecting the consumer much more.
Maybe in your region of the USA chip and signature is prevalent but in my experience (SF bay area) every chip transaction has required a PIN.
Or perhaps it's your card-issuing bank?
AFAIK ALMOST NO US-based banks issue chip+pin credit cards. Only chip+signature. Debit cards are indeed chip+pin, but in USA you'd have to be careless to ever use a debit card anywhere but an ATM.
Chip+PIN doesn't really exist for US CCs. I've setup PINs on a few cards but they only work overseas where the terminal refuses signature.
I haven't seen anyone or myself swiped in about 10 years, and that was only because the chip reader wasn't working at the time.
A lot of this has to do with Quick Chip, which forgoes writing the host authorizer ID to the credit card's chip thus simplifying and shortening the transaction. QC rollouts happen at whatever pace retailers want, so you'll see some shops using QC and others that don't.
https://www.cardfellow.com/quick-chip-slow-chip-card-transac...
Having a requirement that all vendors use chip-enabled cards simply increases the cost of producing each cloned card; so the goal is to raise the cost of committing the fraud.
At least that is my understanding.
Just because you are using the chip doesn't mean you are doing an EMV transaction. The unique transaction codes only happen with and EMV transaction, almost every time you dip your card it's a regular old card transaction, just as if you swiped the card. Why?
Getting EMV certified requires every part of the transaction chain to go through unified system testing, for each combination of hardware, software, card type, processor, issuing bank.
I've been eyeballs deep in this nonsense for the last year or so. We just can't justify the expense of getting EMV certified, so we just accept the chip and do a regular transaction.
As a consumer you have no way of knowing if your transaction is and EMV transaction or just a chip enabled regular transaction.