[148638] in cryptography@c2.net mail archive
[Cryptography] HSBC's Password Approach: Impressive
daemon@ATHENA.MIT.EDU (Kent Borg)
Mon Dec 23 18:12:52 2013
X-Original-To: cryptography@metzdowd.com
Date: Mon, 23 Dec 2013 12:00:34 -0500
From: Kent Borg <kentborg@borg.org>
To: cryptography@metzdowd.com
Errors-To: cryptography-bounces+crypto.discuss=bloom-picayune.mit.edu@metzdowd.com
I have an account with one of HSBC's banks and in my case their password
use is pretty impressive. (I don't know how it varies around the world
with different instances of HSBC.)
1. They chose the password, not me. (So not duplicated elsewhere, at
least not initially.) Short, alpha-numeric.
2. When I login, they want my account number. Their Javascript doesn't
allow pasting, I used to use X11 middle-click pasting but one day they
decided I was a specific piece of Windows malware and locked me out
until I had my computer professionally cleaned. I managed to talk them
out of that for my Linux machine--but I don't middle click any more. (I
actually ended up talking to someone pretty real.) They are watching
their attacks, maintaining a security model on each customer.
4. They want the answer to a security question (effectively a password I
have chosen).
5. They want their password--but only a few positions, they show a form,
graying out boxes for positions they are not interested in.
In their model they are keeping track of when they have used which
character positions (when a keyboard sniffer might have discovered a
position). They seem to settle on the same character positions for a
long time, until something interesting happens (such as logging in with
a smart phone app), then they shift them. I bet their HTML doesn't
obviously reveal which positions they are requesting, to make the
sniffing harder.
They are being pretty clever to make up for terribly endpoint security.
-kb
_______________________________________________
The cryptography mailing list
cryptography@metzdowd.com
http://www.metzdowd.com/mailman/listinfo/cryptography
