ietf
[Top] [All Lists]

Re: We need an architecture, not finger pointing.

2015-10-28 10:20:18
On Oct 27, 2015, at 2:17 PM, Christian Huitema 
<huitema(_at_)microsoft(_dot_)com> wrote:
Identity checks matter. Lots of the discussion focused on SPAM, but the 
"acute problem of the day" is actually phishing, and specifically forging a 
mail that appears to come from someone you trust, to entice you to open a 
document or visit a URL that you should not. That's a pretty common step in 
the chain of events that leads to another "42 million user accounts 
compromised in a breach." 

This is correct.   If I seem a bit strident on this issue, it’s because I’ve 
actually had a family member personally affected by this problem, in a rather 
severe way.   The ability to verify that mail actually came from whom it claims 
to have come is quite important particularly for aging family members who might 
not not be as good at detecting scams as they once were.   And for that matter, 
I’ve been momentarily fooled once or twice in recent years—the amateurs give us 
a false sense of security, but some phishers are _very_ skilled.

Of course, part of this is a UI issue, which is out of scope, but perhaps worth 
mentioning: MUAs should never present a blinded URL.   If the HTML looks like 
this:

<a href="http://nefarious.example.org/QOJWEOJOWJCJ#UR1OJOJFOIJ?hack=yes 
<http://nefarious.example.org/QOJWEOJOWJCJ#UR1OJOJFOIJ?hack=yes>">http://www.example.com/
 <http://www.example.com/></a>

The user should see this:

http://nefarious.example.org/QOJWEOJOWJCJ#UR1OJOJFOIJ?hack=yes 
<http://nefarious.example.org/QOJWEOJOWJCJ#UR1OJOJFOIJ?hack=yes>

I mention this only because PHB brought up architecture and Christian mentioned 
enticing people to visit URLs.   The definition of "working" has to include not 
being subject to obvious UI vulnerabilities that are only safe if the end user 
is Bruce Schneier.   Perhaps we need a Consumer Reports for commonly-used 
software.