search

camaraproject / QualityOnDemand

Repository to describe, develop, document and test the QualityOnDemand API family
https://wiki.camaraproject.org/x/zwOeAQ
Apache License 2.0
42 stars 59 forks source link

Improvement proposal for QoSProfile_Mapping_Table.md #62

Closed tlohmar closed 1 year ago

tlohmar commented 1 year ago

The meaning of the different QoS Profile Labels are pretty unclear for a developer, who has no insights into the 5G Network realization (e.g. the meaning of 5QI ==6)

The table is extended with two examples, describing the meaning of the QOS Profile Labels and also constrains, in which cases the Network Service cannot be provided.

hdamker commented 1 year ago

Call on 2022-11-04: comments to be addressed, PR will fix #61

eric-murray commented 1 year ago

@tlohmar Having thought about this a bit more, I would suggest to update "statistically assured" to "statistically assured according to the agreed SLA".

My thought is that somehow we have to agree with the API caller what the different QoS levels mean contractually, and I'll call this an "SLA". This would explain that the minimum throughput is not guaranteed under any and all conditions, but only for a specified (high) percentage of them. Quite how this SLA is presented to and accepted by the API caller is an open issue, but to have a contract, it must exist.

Other than that, I'm happy with the text and Herbert's proposed changes.

hdamker commented 1 year ago

There are no further comments since two days, so will merge it.

ToshiWakayama-KDDI commented 1 year ago

Sorry for being late, but I have some comments to this and would like to discuss during today's call. Thanks.

jlurien commented 1 year ago

In TEF we are also discussing internally this issue and would like to comment on it. Sorry, not aware of a deadline.

ToshiWakayama-KDDI commented 1 year ago

Hi, I raised some issues on Mapping table at the last QoD API meeting on 18th Nov, and it is supposed to be discussed by email discussion. I would like Herbert to start email discussion on the Mapping table, or if not, I would propose to re-open this PR and discuss it. Thanks.