Closed italobusi closed 1 year ago
This was the proposal in issue #1 to manage the position within the NE:
+--ro position
| +--ro ne-name string
| +--ro rack? uint8
| +--ro shelf? uint8
| +--ro sub-shelf* uint8 // when ../shelf
| +--ro slot? uint8
| +--ro sub-slot* uint8 // when ../slot
| +--ro port? uint8 // when (not ../slot and not ../shelf and
| // not ../rack) or (../slot)
2022-05-25 Network Inventory weekly call
2022-05-25 Network Inventory weekly call
- [ ] @aguoietf , @sergiobelotti , @jbouqui153 : provides a proposal for the location/position information, taking into account that it should be generic and not applicable only to optical NEs
OpenConfig has a string position but it seems to exploit on parent and child pointer (leafref) to show the structure of all the components in the node and a the containments.
2022-07-06 Network Inventory weekly call
- @aguoietf , @sergiobelotti , @jbouqui153 : provides a proposal for the location/position information, taking into account that it should be generic and not applicable only to optical NEs
The proposal for IETF 114 is to replace the location container with a free-form string:
OLD
+--ro location
| +--ro equipment-room-name? leafref
| +--ro ne-name? leafref
| +--ro rack-number? leafref
| +--ro chassis-index? uint16
| +--ro slot-index? uint16
| +--ro port-index? uint16
NEW
+--ro location string
The reason is that a free-form string has more flexibility than a pattern like "/ne=\
The need to report the location information versus inferring it from the parent information (to updated with some index value) should be further discussed (e.g., during the presentation at IETF 114)
2022-11-23 Network Inventory weekly call
Chaode: For issue #3, we all agreed to provide a string format for the components' location. And for IP technology, if the there are some other type component, we just need to modify the pattern which could be compatible. And for issue #41, we have actually provide UUID information in the parent-references structure, so a uuid-based location is not needed.
Jeff: yes, the uuid-based location is not needed. And for the equipment room's name, i remember we have define it in the someplace.
Chaode: The equipment room's name is the location of network element.
Issue #41 can be closed
2023-02-15 Network Inventory weekly call
Since the location is a free form string it would be worthwhile documenting some format constraints, similar to those defined in ONF TR-547, in the I-D
This is a common format used in Optical networks so it is recommended at least for Optical networks but it seems also applicable to IP and MW
This format can be described as a "recommended" format for Optical, IP and MW, unless there are objections from IP and MW experts
2023-06-28 Weekly Call The text has been provided in the YANG file, we agree to close this issue.
To be fixed by PR #98
Check whether uint8 is sufficient for rack/shelf/slot/port identification or a longer unsigned integer type should be used instead