Re: [lmap] draft-ietf-lmap-information-model-05: Controller timeout

["Carey, Timothy (Timothy)" <[e-mail]>]

Sorry - For the range I was thinking 16-bit range not 32-bit. 32-bit should be fine. My mistake.

I will say though that the interface speed may become a problem if you leave it as an int since the unit is bps.

-----Original Message-----
From: Carey, Timothy (Timothy) 
Sent: Tuesday, May 12, 2015 6:59 AM
To: 'Juergen Schoenwaelder'
Cc: [e-mail]
Subject: RE: [lmap] draft-ietf-lmap-information-model-05: Controller timeout

Juergen,

So then we are saying that schedules (I think you mean scheduled tasks?) are operationally disabled and are reenabled with the controller is contacted.

We that would mean all scheduled tasks that are not related to controller communication - right? Since everything is a scheduled task now?

As to the int vs unsignedInt - I don't necessarily care except that one might ask what a meaning is for negative values unless you use lower threshold constraint of 0. Also you lose have the range of the timeout. What is the size of the int - 32 or 64-bit? My point is that if it is 32bit you its timeout would be limited to like 9+ hours (32767 seconds)

-----Original Message-----
From: Juergen Schoenwaelder [mailto:[e-mail]]
Sent: Tuesday, May 12, 2015 5:07 AM
To: Carey, Timothy (Timothy)
Cc: [e-mail]
Subject: Re: [lmap] draft-ietf-lmap-information-model-05: Controller timeout

On Mon, May 11, 2015 at 07:08:47PM +0000, Carey, Timothy (Timothy) wrote:
> Team,
> 
> I was reviewing draft-ietf-lmap-information-model-05 and had a 
> question on the
> ma-controller-lost-timeout:  A timer is started after each successful contact with a controller.  When the timer reaches the controller-lost-timeout, all schedules will be disabled.
> 
> We changed this from a threshold to a timeout but the data type is still an <int> and there are not units assigned to the timeout (e.g., milliseconds, seconds).
> 
> Also the description is vague - When we mean schedules disabled - so we mean administratively or operationally? In other words - does it take an explicit action from the controller to reenable the schedule or simply the next contact. Also is this the schedules that are disabled or scheduled tasks.
> 
> I would assume the timeout would be in seconds and we should change the data type to unsignedInt unless of course a negative number has some meaning?
> 

I agree this is under-defined. We generally have only used 'int' and we did not make a distinction between int and unsigned so I would leave that alone (or we have to make changes in multiple places).

What about this rewrite:

   ma-controller-lost-timeout:  A timer is started after each successful
                             contact with a controller.  When the timer
                             reaches the controller-lost-timeout
                             (measured in seconds), all schedules will
                             be disabled, i.e., no new schedules will be
                             executed (and hence no new tasks started).
                             The disabled schedules will be reenabled
                             automatically once contact with a
                             controller has been established
                             successfully again.

/js

-- 
Juergen Schoenwaelder           Jacobs University Bremen gGmbH
Phone: [phone]         Campus Ring 1 | 28759 Bremen | Germany
Fax:   [phone]         <http://www.jacobs-university.de/>