Re: [Teas-ns-dt] Availability

I agree with both points, and this is what the document meant to say, SLA = Σ(SLO1 && SLOn). I’ll add some text
Every SLA is slice specific, user is only interested if their SLOs are met, anything outside of this particular set of SLO’s is irrelevant.

Cheers,
Jeff
On Jun 11, 2020, 7:51 AM -0700, Jari Arkko <jari.arkko=40ericsson.com@dmarc.ietf.org>rg>, wrote:
>
> > I also think that the service availability should be based on the conformant of all SLOs listed in the SLA. Failure in meeting any one of the SLOs should be considered unavailable for the service.
>
> I agree.
>
> > And in addition to the failure or outage case, some SLO of one network slice may be impacted by the misbehavior of traffic in other network slices. Thus to meet the required availability, the impact between slices also needs to be taken into consideration.
>
> The impact of any misbehaviour, accidents, or any problem to keep enough capacity to satisfy committed contracts are all reasons why there's a failure, and consequently, reduced availability.
>
> But I don't think it matters in the metric what the reason for the failure was.
>
> Jari
>
>
> --
> Teas-ns-dt mailing list
> Teas-ns-dt@ietf.org
> https://www.ietf.org/mailman/listinfo/teas-ns-dt