[Nml-wg] Lifetime

[Aaron Brown]

Hey Jeroen,

On Mar 13, 2013, at 10:01 AM, Jeroen van der Ham <vdham at uva.nl> wrote:

> Hi,
> 
> Time is always a very difficult problem. We have tried very hard to define this properly in the schema. There was one small gap that we missed: when a Network Object did not have a Lifetime *and* no new Topology object with a version attribute was published, or a new one is and does not contain a description of that Network Object.
> 
> I've expanded the description in the schema to this, can you please check if this covers all cases and whether you agree with it?
> 
> 
> If a Network Object has no associated \emph{Lifetime} objects, or the start or end attribute of a Lifetime object is missing, the default lifetime may be assumed to start on or before the time specified in the version attribute of the most specific Topology object that contains this Network Object. The end of that assumed lifetime is indefinite, until a Topology object with a higher version number is published. This new description can define a new Lifetime for the object, or the Topology. If the new description does not contain the Network Object, the end time is assumed to have passed.

What does "most specific Topology object that contains this Network Object" mean? What happens if the same element is referenced by 2 Topologies with different lifetimes e.g. version 1 of the Topology has node A, and version 2 of the Topology still has node A, but only version 2 has node B. What lifetime does node A inherit?

Beyond that, I'd say that there is no default. If a start or end time is unspecified, it's unknown when the element came into existence nor when it goes out of existence. Folks can then interpret that however they like.

Cheers,
Aaron

> 
> 
> Jeroen.
> 
> _______________________________________________
> nml-wg mailing list
> nml-wg at ogf.org
> https://www.ogf.org/mailman/listinfo/nml-wg
>