[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]


To: EPP Provreg <ietf-provreg@cafax.se>
From: Patrick Mevzek <provreg@contact.dotandco.com>
Date: Tue, 14 Apr 2009 15:11:05 +0200
Content-Disposition: inline
In-Reply-To: <046F43A8D79C794FA4733814869CDF07029FD729@dul1wnexmb01.vcorp.ad.vrsn.com>
Sender: owner-ietf-provreg@cafax.se
User-Agent: Mutt/1.5.18 (2008-05-17)
Subject: Re: [ietf-provreg] AD Review Comments:draft-hollenbeck-rfc4934bis-00

[about TLS & close_notify]

Hollenbeck, Scott <shollenbeck@verisign.com> 2009-04-14 14:14
> I'm reasonably certain
> that this is behavior that's typically buried in a toolkit instead of
> being implemented by an EPP developer, so perhaps that new text should
> be removed from the document.
 
I agree with you on that.

As James, I'm using a toolkit to deal with SSL stuff (a wrapper
around openssl called IO::Socket::SSL under Perl), 
I do not know if it does a TLS_close_notify.
Per its documentation, it does do it properly, even warning not to do
a shutdown before a close has it needs to send this TLS notification
on proper shutdown.

Also, most of the time, the connection is first closed by the
server, after the client <logout/> message.
A quick ssldump gives me:

1 8  0.2824 (0.0000)  S>CV3.1(992)  application_data
1 9  0.2967 (0.0142)  C>SV3.1(768)  application_data
1 10 0.3632 (0.0664)  S>CV3.1(576)  application_data
1 11 1.2652 (0.9020)  C>SV3.1(336)  application_data
1 12 1.3261 (0.0608)  S>CV3.1(368)  application_data
1 13 1.3261 (0.0000)  S>CV3.1(32)  Alert
1    1.3261 (0.0000)  S>C  TCP FIN
1 14 1.3275 (0.0013)  C>SV3.1(32)  Alert
1    1.3279 (0.0004)  C>S  TCP RST

The alerts are probably the close_notify.

-- 
Patrick Mevzek
Dot and Co <http://www.dotandco.com/> <http://www.dotandco.net/>

Home | Date list | Subject list