Hacker Newsnew | past | comments | ask | show | jobs | submitlogin
2011: The year your mobile phone becomes your wallet (cnn.com)
21 points by gaiusparx on Jan 19, 2011 | hide | past | favorite | 22 comments


How secure is this? Waving a device near something to complete a money transaction seems too easy.

At Defcon 18, there was a presentation on how to access RFID from long range (http://www.youtube.com/watch?v=nEBrslz0Xf4). It makes me nervous that other people can also do this stuff.

There are already cards that do NFC, such as MBTA (public transportation) Charlie cards and American Express credit cards. However once NFC becomes even more common, will this invite the bad kind of hackers to bump up their efforts to steal your money?

So, how secure is this?


a phone can show you a "confirm this payment" prompt though, which i imagine puts it a step above paypass/oyster card type pieces of plastic.


A phone can also have a virus or trojan horse app on it. Even easier to take your money if you can get access directly on the phone. Forget radio waves, just get the user to install an app on there that will transfer funds over.

I think all of the smart phone operating systems have had a trojan horse app on them at some point. Most of them probably just did simple things before like a backdoor, making phone calls or stealing info. If your phone also controls your money...


Considering that traditional credit card security is based on just knowledge of simple shared secret - (card number, name, expiry date, cvv2) 4-tuple, and considering this secret is shared with everyone you do business with, this is still certainly more secure.

Also, trojans still have to compromise the OS security. It's not the same deal as with non-secure information (like spyware games accessing the contacts list, just because users don't care about app permissions) - the certificate store will be certainly guarded way more carefully.


With more users 'jailbreaking' or 'rooting' their phones, it's not as 'secure' as before. The point of doing those things is to break the phone's security and allow you to do more things. What if the creator decided to put in a little extra code.. or it introduced more vulnerabilities..

* I'm not trying to make you guys paranoid. Just wondering about the possibilities and playing devil advocate.


Well, not exactly, "jailbreaking" is about user gaining access to a phone, not applications. I don't know the details (I don't own iPhone or Android device, and N900 comes with root access almost out-of-box), but I believe pure jailbreaking shouldn't hurt OS security mechanisms too much. I may be wrong, though.

However, you have a point: classic "dancing bunnies" problem[1] is certainly out there. Promise average user nice things, ask him to do some cryptic actions, and he'll happily follow your instructions without really understanding any consequences.

[1]: http://en.wikipedia.org/wiki/Dancing_bunnies


Drdaeman, it opens up the possibility of such "hostile" data access. For an application to simply start accessing random data in such a way that it could "access" financial details or actually interact with a payment provider... it would have to explicitly be given root permission. I give very few applications root. Terminal gets it, Rom Manager gets it. That's it.

A random app from the Market? Hell no, it's not getting root and it's not getting access to my data.

The thing is, this vector already exists. Download the Paypal app. Login. Now download drivebyacct2's malicious app on the Market. If you grant it root access through SuperUser, it could theoretically read out your Paypal credentials and do all kinds of stuff.

Personally, I consider it to be very low risk.


Don't know how secure it is in this implementation but it should be easy to implement something that requires you to cryptographically sign the transaction.


Would have to be a rolling-key, I'm guessing...but as I mentioned in another comment, the interval would be the kicker. Maybe it's simpler, now that I think of it. A rolling-key system with 2-way authentication, a one-time key becoming invalid after a transaction is made.


Couldn't you just type in a passphrase on your phone and have it sign the transaction?


Whatever solution is created, I don't want to have to manually verify every transaction.


For the past five years, I have been paying parking in my city with my mobile phone using a simple SMS without any transaction fee. A small city of 20000 inhabitants lost in the former East Germany.

I suppose 2011 is more the year where you will pay $5 for your Latte with your mobile phone on the condition to have an account with an organization charging you a percentage when loading your account, plus a small transaction fee and doing the same to the shop using this system.

I am sarcastic, maybe the effect of going through a PCI DSS compliance extortion scheme yesterday. Out of topic technical bonus point, the SSL ciphers allowed for your server to be "compliant": !EXPORT:!eNULL:!MEDIUM:!LOW:TLSv1:SSLv3.


It's never going to be as entertaining as when your wallet becomes your mobile phone:

http://www.merlinmann.com/phoneguy/


2011 may be a big year for changes in the way we bank, with BankSimple releasing and this technology finally making its way to the US.

It's no secret that the banking industry is screwed up. It will be interesting to follow how this industry changes as new startups and technologies are applied to traditional banking models. Hopefully somehow I end up with more money in the process :)


My understanding of mobile phone security is that it's somewhere between abysmal and non-existent.


Almost 10 years ago: http://press.nokia.com/PR/200109/834842_5.html

Is 2011 also the year of Linux on desktop?


Soon there will be more smart phones than desktops, and most of those run Linux, so the desktop has become irrelevant :-)


The thing is how secure is any of our payment methods today? Can't credit cards be replicated as well?


Yeah, but the attacker needs access to the magnetic sensor in order to install a ripping device. Physical security prevents it in most cases. In the case of a phone, how could you prevent playback attacks? Say I walk up to a Coke machine and authorize a $1.00 payment, and someone nearby is able to capture that protocol stream. All they'd have to do is play it back to the Coke machine after I leave, and at that point they are welcome to unlimited corn syrup swill. I'm no cryptographer, but I can't think of any way to mitigate it. Something like SecurID, whereby you are given a new token code every minute, might work, but the intervals for new codes would have to be tighter, unless you plan to stand around and monitor everyone else's purchases for the entire interval. This is one for the really smart guys to figure out, and I expect that someday, it will become reality. Can't wait.


Unfortunately we'll still have to carry our wallets until our ID is on our phone as well.


No thank you.


I use Venmo - I love Venmo - I can't wait for it to become bigger.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: