Agreed to dedicate one week / 4 days / two hours per day for the review of TR-547: July 5,6,7,9 10-12 CEsT
Next TAPI weeky call - 2 hours
|10 mins||Inter-SDO updates|
Ramon Casellas and Arturo presented the TAPI extensions for GNPY to TIP MUST, 40' high level overview.
Arturo, apparently there is already an active liaison between ONF and TIP, we need to check internally to ONF.
Nigel Davis Nokia presented TAPI to O-RAN.
Technical discussion on Integration of Streaming and Fault Management
Andrea Mazzini presents the proposal for common alarm/TCA data structure to be used in both Notification and Streaming contexts.
The AlarmInfo and TcaInfo are deprecated, their content is modeled by the DetectedCondition class, which will augment either Notification signal (picture above) or ConditionDetector (picture below).
Nigel Davis keep exploring a formal way to model the "changed attributes", i.e. to avoid name-value pairs.
Andrea Mazzini presents a reorganization of Notification signal:
Agreed to keep only ALARM and TCA as ConditionType enumeration entries.
Agreed to add FLEETING to SimpleDetectorState enumeration. It is the delta function, e.g. when an event has a very short life (rapid Active-Clear cycling), hence is notified/streamed after its occurrence.
Noted that ACTIVE_NO_EXPLICIT_CLEAR applies to PM metrics which can only increase (counters), hence the "clear" criteria is conventionally the end of a measurement period.
Nigel Davis points out that the severity is always insufficient to address the problem, because:
The overall approach is preliminary agreed:
GNPy/CCAMP – TAPI alignment, presentation of candidate TAPI extensions – continued
Andrea Mazzini shows some updated pictures:
Discussion on capability vs. operational model:
An analogous approach shall be considered for amplifiers, i.e:
Huy Tran clarifies that currently the GNPy process is considering only the operational values of amplifiers.
It is necessary to clarify which is the CCAMP model of amplifier capabilities. Similar consideration for transceiver, its capability and operational model.
Ramon Casellas it is assumed that a bidirectional NEP has symmetric spectrum support, otherwise the unidirectional model shall be implemented.
Huy Tran confirms that current GNPy application does not differentiate between different load conditions. For future consideration.
Use Case 0d – clarification on SAPI/DAPI vs. TxTI/ExT
Andrea Mazzini asks for confirmation of last week agreement:
Model the NEP PlugId attribute as:
Considering OTN technology:
The solution is confirmed.