Re: Shutdown

Richard_R.Moore (MOORERR@msu.edu)
Fri, 08 Aug 97 12:01 EDT

> Date: Fri, 8 Aug 1997 09:26:14 -0400 (EDT)
> From: Shawn <shawnc@ime.net>
> To: MegaZone <megazone@livingston.com>
> cc: portmaster-users@livingston.com, rfe@livingston.com
> Subject: Re: Shutdown (fwd)
> Reply-To: Shawn <shawnc@ime.net>
>
>
>
> On Thu, 7 Aug 1997, MegaZone wrote:
>
> > Busies out modems as they drop, and doesn't accept any new connections
> > (returns busy). Reboots when all connections are gone.
>
> So users get busy signals for 8 hours until everyone has dropped off?
>

YES! If that is what it takes to not abruptly cut users off. Usually, there
is some excess capcity in your dailup so that not everyone will get busy
signals. Of course, we should also have the option to just abruptly shut
things down.

Consider this senario: I have a problem that will require shutdown of a PM3,
but it is not catastrophic. My peak usage end at 2300. So, at 2330, i tell
the PM3 not to answer any more calls. Between 2330 and 0600 i can easily do
without 46 (or even sixty) lines. At 0600, perhaps there is still one or two
hangers on -- i then tell the PM3 to terminate all connections. I do what
ever maintenance i need to do and have the PM3 back up by 0800. Mission
accomplished with very little customer aggravation.

Contrast this with the current, abismal situation: My dial-in service is on a
simple trunk hunting scheme. The PM3 i want to turn off is the one handling
the first set of lines. Even at 0300, i will have at least 46 active users.
Therefore i HAVE NO CHOICE but to abruptly terminate these 46 users because
livingston will not give me the simple option of idling down the box.

It is only "reasonable design" to give the option to not answer any future
calls. I was VERY surprised to learn that it does not!