[Cryptography] So please tell me. Why is my solution wrong?
Joseph Kilcullen
kilcullenj at gmail.com
Thu Feb 16 16:58:08 EST 2017
<html>
<head>
<meta content="text/html; charset=utf-8" http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">On 16-Feb-17 6:28 PM, Tom Mitchell
wrote:<br>
</div>
<blockquote
cite="mid:CAAMy4UR3bZAFxMWFhwD=pk3KPXP93wALyoqpam9G4QoWA-emYA at mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote"><br>
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">
The solution in 1 sentence reads:<br>
<br>
Once you view the web browser as an actor in the
cryptography protocol everything else is classic
cryptography i.e. your browser must authenticate itself by
presenting a shared secret. That's it!</blockquote>
<div><br>
</div>
<div>The solution is not a general case solution. It helps
only in the special case of a hijacked web environment. <br>
It does not catch the case of a transparent application
that can intercept keystrokes or even a paste <br>
from a previously filled mouse copy buffer. <br>
<br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
True. I'm focusing on phishing which tries to bypass TLS. If your
computer has been hacked in any way, virus etc. this solution won't
save you.<br>
<br>
<br>
<blockquote
cite="mid:CAAMy4UR3bZAFxMWFhwD=pk3KPXP93wALyoqpam9G4QoWA-emYA at mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div>Transparent applications must ........ a sense of
security.<br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
Agreed.<br>
<br>
<blockquote
cite="mid:CAAMy4UR3bZAFxMWFhwD=pk3KPXP93wALyoqpam9G4QoWA-emYA at mail.gmail.com"
type="cite">
<div dir="ltr">
<div class="gmail_extra">
<div class="gmail_quote">
<div><br>
It still has value but is not sufficient. If local
javascript can see the local image then<br>
it can be stolen or used in place.<br>
<br>
</div>
</div>
</div>
</div>
</blockquote>
<br>
True. In fact the image was chosen as a good example of a 'web
browser' - 'user' shared secret. The idea of javascript grabbing the
image totally undermines the idea. So my preference is that it be
built into browsers. That the developers understand what's
happening.<br>
<br>
<br>
**********<br>
<br>
An addition to the one line description:<br>
<br>
One line: Once you view the web browser as an actor in the
cryptography protocol everything else is classic cryptography i.e.
your browser must authenticate itself by presenting a shared secret.<br>
<br>
Browsers have two personalities. (A) installed software which can
access the hard disk and (B) the browser canvas area which is
controlled by a remote website. This personality cannot access the
hard disk.<br>
<br>
TLS is implemented by personality (A). Without authentication you
don't know which personality you are talking to. When Mallory
utilises personality (B) to counterfeit a website (and/or browser
interface) that's a phishing attack. All that 'full screen
counterfeiting' stuff indicates just how far personality (B) can go.
Hence on start-up of TLS our browser must authenticate itself as
personality (A). Personality (B) is not able to present the correct
image in fig1.<br>
<br>
This goes to the heart of the research i.e. I sought game theory
screening strategies to elicit the genuine identity present. TLS
(personality (A)) can show the image. Mallory and personality (B)
cannot.<br>
<br>
Thanks<br>
Joseph<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<br>
<p><br>
</p>
</body>
</html>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20170216/aa953f4f/attachment.html>
More information about the cryptography
mailing list