The oldest two mechanisms of authenticating on credit cards.
From oldest to newest, they are:
Printed data on card.
Magstrip (which basically has the same data in machine-readable form).
Smartcard chip with contacts.
Wireless.
The first two mechanisms hand over all the data required to impersonate the cardholder whenever used, which isn’t very secure. Yes, there’s value to keeping a mechanism around for a while to permit transition time, but we should have had tap-to-pay hardware on PCs and phones and the like a long time ago.
I’d assume so, but more importantly, for both, there’s a cryptographic signature being performed by the card. The credentials never leave the card — there’s a private key on the card, and what goes out is a signature on the transaction, which is useless for doing other transactions.
That’s not true for all cards, at the very least. Skimming wirelessly by RFID is or was a thing. The whole backbone of the credit card system is designed to expect the number.
I disagree that we should have a card reader on our computers for payments.
That is just a way too big of a security concern.
I prefer something like the Swedish system Swish, you have a separate app on your phone where you can send money to friends and family as well as pay for stuff online.
Sadly, while Klarna supports Swish, they require the use of a Klarna account to use it, and since most internet shops in Sweden uses Klarna it limits the ability to use it as I want to.
and with that you need a smartphone, with a google-approved operating system and with it half of the factory bloatware, or otherwise you are barred from paying online, right? that sounds such a good idea.
I said nothing about the OS on the phone, why would you assume that I like Android?
that’s not what I assumed. I assume that this app would only support the 2 most popular mobile platforms, and that on android, as is tradition with payment related apps, it would refuse to work when it detects that your phone’s software has been changed in any significant way.
if Swish and BankID could run on an open mobile plattform, I’d be happy with that.
current trend is to make these apps OWASP compliant, which dictates that all apps should at least be an undecipherable, obfuscated black box, and better even make use of the OS’s integrity checking system, like play integrity on android.
My point it to separate the main computer from the payment system while still being convenient.
I am a bit confused as how you missed that…
I did not miss that. I was commenting on this, why it would be harmful in today’s world.
Something attached to the main computer, but with its own firmware/controls is still far better than having no device at all, and relying on external code for verification. Would a discrete box separate from everything else be better (independent of mobile phones as well)? Sure. But a great step that would be progress compared to the current status quo is what the other poster describes, with logic and chip verification running on a device attached to the device or computer with which you wish to pay.
The oldest two mechanisms of authenticating on credit cards.
From oldest to newest, they are:
Printed data on card.
Magstrip (which basically has the same data in machine-readable form).
Smartcard chip with contacts.
Wireless.
The first two mechanisms hand over all the data required to impersonate the cardholder whenever used, which isn’t very secure. Yes, there’s value to keeping a mechanism around for a while to permit transition time, but we should have had tap-to-pay hardware on PCs and phones and the like a long time ago.
Does smartcard and wireless actually have an encryption layer of some kind?
I’d assume so, but more importantly, for both, there’s a cryptographic signature being performed by the card. The credentials never leave the card — there’s a private key on the card, and what goes out is a signature on the transaction, which is useless for doing other transactions.
That’s not true for all cards, at the very least. Skimming wirelessly by RFID is or was a thing. The whole backbone of the credit card system is designed to expect the number.
I disagree that we should have a card reader on our computers for payments.
That is just a way too big of a security concern.
I prefer something like the Swedish system Swish, you have a separate app on your phone where you can send money to friends and family as well as pay for stuff online.
Sadly, while Klarna supports Swish, they require the use of a Klarna account to use it, and since most internet shops in Sweden uses Klarna it limits the ability to use it as I want to.
and with that you need a smartphone, with a google-approved operating system and with it half of the factory bloatware, or otherwise you are barred from paying online, right? that sounds such a good idea.
no.
I said nothing about the OS on the phone, why would you assume that I like Android?
I am an iPhone user, but that is beside the point, if Swish and BankID could run on an open mobile plattform, I’d be happy with that.
My point it to separate the main computer from the payment system while still being convenient.
I am a bit confused as how you missed that…
that’s not what I assumed. I assume that this app would only support the 2 most popular mobile platforms, and that on android, as is tradition with payment related apps, it would refuse to work when it detects that your phone’s software has been changed in any significant way.
current trend is to make these apps OWASP compliant, which dictates that all apps should at least be an undecipherable, obfuscated black box, and better even make use of the OS’s integrity checking system, like play integrity on android.
I did not miss that. I was commenting on this, why it would be harmful in today’s world.
Something attached to the main computer, but with its own firmware/controls is still far better than having no device at all, and relying on external code for verification. Would a discrete box separate from everything else be better (independent of mobile phones as well)? Sure. But a great step that would be progress compared to the current status quo is what the other poster describes, with logic and chip verification running on a device attached to the device or computer with which you wish to pay.