[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: CAP: autologout
The issue of having a NOOP command, and allowing the server to ignore
it did come up in January 2000 (see the archive).
Looking at the archive, it was agreed to allow the server to ignore
NOOP and still drop the connection.
What you are suggesting is to return an error rather than simply
NOOP. Should we return an error, or simply allow the server to drop the
connection despite the NOOPs?
The reasons NOOP was kept was:
- allow the client to attempt to keep the connection up
- a simple way to check if the server is still alive, Like PING.
(In fact, another issue that came up in the NOOP thread was if
NOOP should be renamed to "PING.")
However, we should have a way of discouraging avoid the situation
a CUA calls NOOP often.
In section 4.4 we have:
"A CUA may be discouraged from attempts to do useless things only to
Yet, we also give them NOOP that may encourage CUAs to do such useless
That part of section 4.4 can be interpreted that the CS can terminate
the connection or return some error code if the CUA calls NOOP too many
John Stracke wrote:
> A couple of questions about autologout on idle:
> * 4.4: "When a timeout occurs, the server drops the connection to
> the CUA". Should there be some sort of notification first? We've
> got a notification for server shutdown (difficult as that may be
> to implement); one for idle timeout seems reasonable.
> * 7.1.12: the NOOP command seems like a way of evading the server's
> idle timeout policy. Many servers I've known with such timeouts
> don't want them evaded (for example, I had an ISP which had an
> idle timeout on their PPP connections, and warned that ping
> packets didn't count, because people were using ping to keep the
> connection alive). Maybe there should be an error code reporting
> that the NOOP has been ignored?
> |John Stracke | http://www.ecal.com |My opinions are my own.|
> |Chief Scientist |=============================================|
> |eCal Corp. |Please do not adjust your mind--reality is |
> |francis@xxxxxxxx <mailto:francis@xxxxxxxx>|malfunctioning. |