<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN">
<html>
<head>
<meta content="text/html;charset=UTF-8" http-equiv="Content-Type">
</head>
<body bgcolor="#ffffff" text="#000000">
Hmm.... I'll have to check this out.<br>
<br>
anorexicpoodle wrote:
<blockquote cite="mid:1257978568.2377.61.camel@poodle-x300" type="cite">
<pre wrap="">Actually I would be far more concerned about the spoofing of refer
messages, which is one of the message types it doesn't auth. It struck
me as unusual that it would auth a bye and not a refer.
On Wed, 2009-11-11 at 16:20 -0600, Lee Riemer wrote:
</pre>
<blockquote type="cite">
<pre wrap="">I agree. Shouldn't be too hard if you're sniffing the line, say on a
wi-fi connection. Listen for INVITEs, pull dialog info, spoof CANCELs.
Nice little DOS attack.
Alex Balashov wrote:
</pre>
<blockquote type="cite">
<pre wrap="">Lee Riemer wrote:
</pre>
<blockquote type="cite">
<pre wrap="">What's wrong with authenticating BYEs and CANCELs? I see them as
just as important as an INVITE. Anyone can some around, sniff the
dialog and spoof a BYE or CANCEL.
</pre>
</blockquote>
<pre wrap="">They'd have to spoof a Call-ID, From and To tags, correct CSeqs, any
loose-routing parameters present in the Route and/or Record-Route
headers, and the Route header itself.
I suppose it *could* be done... but...
</pre>
</blockquote>
</blockquote>
<pre wrap=""><!---->
</pre>
</blockquote>
</body>
</html>