OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

obix-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: Response to comments cisi@ix.netcom.com 2


The alarming model is actually pretty simple if you view it as simply a
mixin of three things:

  - when the alarm happened - an edge event captured by obix:Alarm

  - when the alarm condition ceased to happen (if it makes sense) -
captured by obix:StatefulAlarm

  - when the alarm was acknowledged (if it makes sense) - captured by
obix:AckAlarm

 

I think understanding that not all alarms need to be acknowledged is
fairly straight forward and really is just an aspect of some alarms that
occurs independent of the other transitions.

 

But not all alarms are stateful - to take a simple IT example.  A
stateful alarm is "disk full" - that is a condition which occurs during
a period of time and may transition back to normal when space is freed
on the disk.  An example of stateless alarm would be "unauthorized login
attempt" - it's an edge event which occurs at in instance in time -
there is no transition back to normal therefore such an alarm would
never implement obix:StatefulAlarm.

 

The status bits only provide the current snapshot of the 2 variables: is
it currently in alarm and is there an outstanding acknowledgement?  It
is not the intention of status to provide information on the entire
state machine and history of the alarm - that is what the obix:Alarm
itself is for.  Typically status is just used to color code the object
in a graphical display to let an operator know that there going on
behind the scenes.

 

Brian



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]