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


To: "'ietf-provreg@cafax.se'" <ietf-provreg@cafax.se>
From: "Liu, Hong" <Hong.Liu@neustar.biz>
Date: Mon, 12 Aug 2002 15:29:53 -0400
Sender: owner-ietf-provreg@cafax.se
Subject: RE: Response Code 2501

Scott,

I should have clarified better what I meant by "normal" in the previous
email. I would also like to distinguish between session and connection in my
posting, too.

By "normal" operation, I mean the client/server exchange after the session
is established. The issue at hand has to do with session management, rather
than a response to a command. Terminating a session does not necessarily
triggers the termination of the underlying transport connection. I hope that
we can address this issue in EPP without being completely tied to the TCP
mapping.

At present, a client can notify the server to terminate a session via a
<logout> command. However, it is not clear in EPP how a server should
terminate a session on its own end. The use of 2501 at least provides a way
for the server to notify the client that it is closing down the session due
to timeout. The client is not necessarily dead and hung at this point. So it
will be able to retrieve the message from its buffer.

I am looking for a solution to problem of server terminating session
management. I will be happy if you or someone else can give me an
alternative solution.

--Hong

-----Original Message-----
From: Hollenbeck, Scott [mailto:shollenbeck@verisign.com]
Sent: Monday, August 12, 2002 11:58 AM
To: 'Liu, Hong'; 'ietf-provreg@cafax.se'
Subject: RE: Response Code 2501


> While I understand that the normal operating mode for EPP is
> client-initiated command/response, this is a special case 
> where the server
> initiates the action due to non-activity by a client. 
> Otherwise, the client
> will be left without any clue why the connection is gone.

Hong,

It's not just the "normal" operating mode, it's the _only_ operating mode
currently defined.  If the client is dead or otherwise hung, a clue isn't
going to help.

I'm OK with leaving this response code in if it might have some future
benefit (like if someone ever writes a draft describing server message
pushing ;-)), but as things are written currently I'd consider a server that
sends an unsolicited 2501 response to be non-conforming.

-Scott-

Home | Date list | Subject list