[Openswan dev] Opportunistic encryption questions

s baker xiptg2 at hotmail.com
Tue Mar 28 23:04:48 CEST 2006

Hash: SHA1

   I have been doing some reading about Opportunistic encryption, and
it looks pretty promising.  However, for what I would like to do
with it, the current implementation has some limitations.  I was
wondering what some thoughts are about an idea for extension I
   The limitation I am running into is that a static IP and
control of reverse DNS is required to seriously take advantage
of opportunistic encryption, the way it is currently implemented
in openswan.  I believe that the number of people who have
this capability is not going to grow very rapidly in the near
future, while the number of users who have a dynamic IP address
will be where the growth is.
   It seems to me that the power of opportunistic encryption
grows as the user base grows.  In order to take advantage of this,
the more people using it the better.  To have more people
who don't have control of DNS to participate, I think that
some other key distribution capability needs to be present, most
reasonably as a fallback in case a DNS record cannot be found.
   This could be accomplished with a keyserver running on the
receiving node.  When the initiator connects, it would see if
there was a static key, then try reverse DNS, then try a
well known port at the destination to ask for a key.  I recall
seeing that there are key servers already available that could
fulfill this requirement.
   I realize that this is more vulnerable to a man-in-the-middle
attack than a reverse DNS key distribution is.  However, this
risk can be reduced when necessary, by techniques such as
communication of the key via a secondary channel (such as a
GUI running at both ends of the connection, that should display
the same key value).  Key signing or logging could
play a part here as well.  However the strongest argument to
reduce the importance of the man-in-the-middle attack is that
if this capability helps get more users who otherwise would not
be able to participate in opportunistic encryption to do so,
the numbers of connections help provide anonymity.  Along with
the argument that some encryption is better than none.
   So I am interested in where this idea would stand.  Is someone
working on it already?  Is it likely to be added to openswan
if it were developed?

Steve baker
xiptg2 at hotmail.com
Version: GnuPG v1.4.0 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org


More information about the Dev mailing list