Bruce,

While I completely understand the desire to have an easy way for implementations to test TCP connectivity, I do agree with Brian that it's a slippery slope. For implementors trying to rush a product to market (and to a degree isn't everyone?), I would expect that if a "no security" mode is an option in the spec, it will be taken as the path of easiest implementation and any number of warnings will simply be ignored - and as a result you'll have countless insecure RELOAD implementations out there.

Can't a developer simply NOT implement TLS when doing their initial development?

Or perhaps someone could make a "reference test implementation" available for developers to use that supports both "TLS" and "non-TLS" connections? That would give developers something to work with... but yet leave the spec as is so that "conforming" implementations must use TLS.

My 2 cents,
Dan


On Nov 4, 2008, at 7:42 PM, Brian Rosen wrote:

<as individual>
I don't like this idea. I think that you are on a slippery slope and I
think it's not that hard.

I'm not implementing reload at the moment, so I can't speak from experience
in this instance.

Brian

-----Original Message-----
From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of
Bruce Lowekamp
Sent: Tuesday, November 04, 2008 3:34 PM
To: p2psip
Subject: [P2PSIP] adding tcp-test option to reload

There has been significant conversation around whether requiring both
TLS and ICE for a minimally functional reload implementation is too
big of a hurdle during development.  Most developers first implement
something without these two components first regardless, but they need
to solve nodes exchanging identifiers (without TLS) on their own and
their protocols are not interoperable for testing purposes.  The
reload authors would like to propose adding the following text, or
something similar, to introduce a tcp test mode options:

        TCP Test Mode is a transport based on TCP but no security
layer. It SHOULD NOT be used in any production environment as it has many security vulnerabilities. It is meant only as simple test mode to facilitate testing and interoperability before moving to full TLS. When a new TCP session of this type is formed, both ends
        of the connection MUST write their binary Node-ID to the wire
before sending any other messages over the session. This allows both sides to discover the Node-ID of the other side and use this in a similar way to the Node-ID discovered when using TLS or DTLS from the certificate in the TLS handshake. This mode MUST not be
        used unless the configuration for the overlay instance
        specifically allows it.

Bruce
_______________________________________________
P2PSIP mailing list
[EMAIL PROTECTED]
https://www.ietf.org/mailman/listinfo/p2psip

_______________________________________________
P2PSIP mailing list
[EMAIL PROTECTED]
https://www.ietf.org/mailman/listinfo/p2psip

--
Dan York, CISSP, Director of Emerging Communication Technology
Office of the CTO    Voxeo Corporation     [EMAIL PROTECTED]
Phone: +1-407-455-5859  Skype: danyork  http://www.voxeo.com
Blogs: http://blogs.voxeo.com  http://www.disruptivetelephony.com

Build voice applications based on open standards.
Find out how at http://www.voxeo.com/free





_______________________________________________
P2PSIP mailing list
[EMAIL PROTECTED]
https://www.ietf.org/mailman/listinfo/p2psip

Reply via email to