Closed italobusi closed 1 year ago
Network Inventory weekly call (October 13, 2021)
Jeff: the operational state is needed, not the other states
To be further discussed after IETF 112
2022-05-18 Network Inventory weekly call
Contribution from @YuChaode : Attributes.xlsx
2022-05-25 Network Inventory weekly call
[x] @YuChaode : start reviewing the common attributes defined in RFC 8348 and see which ones are applicable to our model
Reviewed table provided by Chaode: https://github.com/italobusi/ietf-network-inventory/files/8772764/Attributes.xlsx
Agreements:
use uuid as the key for the component list and be ready to change it if pushed back by IETF WG discussion;
the name should be a user label (aka alias) with no structure;
need to add a location structure;
[ ] https://github.com/italobusi/ietf-network-inventory/issues/3#issuecomment-1138254025
[ ] https://github.com/italobusi/ietf-network-inventory/issues/41#issuecomment-1138254635
The review of the table will continue offline and at the next meeting.
2022-06-01 Network Inventory weekly call
Continued reviewing table provided by Chaode: https://github.com/italobusi/ietf-network-inventory/files/8772764/Attributes.xlsx
2022-06-08 Network Inventory weekly call
YANG model update presented by Yuchaode. Extensions have been removed as agreed in last call. Rack container outside the rack list. Location information for the rack: equipment-room, row-number and column-number. Additional attributes added for the rack: height, width, depth and max-voltage.
Review of the attributes excel table from Yuchaode. Part number is covered by model name but model-name will be renamed to part-number for sake of clarity.
Discussions if we could have a generic approach for the attributes per component instead of defining subset of attributes for different components. An attribute not supported is not reported by a given component.Agreed to use same attributes schema defined in RFC8348 for all components but then one component instance (e.g. a port) is only reporting some of the attributes that makes sense.
Decision to remove state information (e.g. admin-state, oper-state) because here we are focusing on physical inventory and not logical inventory (RFC8348 is a management model including logical state information).
Agreed that rack-location and ne-location are specific with absolute location information while location shall be kept generic for components under NE indicating relative position of the component with respect to its parent component.
YANG tree updated during the call. Last version will be uploaded in github by Yuchaode.
2022-06-15 Network Inventory weekly call
Continued review and updates to the YANG model proposal by Chaode:
2022-07-06 Network Inventory weekly call
Need to add for each component the parent-rel-pos attribute defined in RFC8348 that could be used to know the position of the component within its parent component.
The analysis has been completed before IETF 114.
2022-09-28 Network Inventory weekly call
[x] @jbouqui153 @sergiobelotti : check the correspondence of RFC component attributes with the OpenConfig model and see if any attribute is missing
2022-10-13 Network Inventory weekly call
Jeff has provided a table to compare the attributes in RFC8348 with the attributes defined in OpenConfig and T-API: Attributes_ Openconfig Jeff.xlsx
The list has been reviewed in detail during the call
Conclusions:
Evaluate whether all the attributes defined in RFC8348 needs to be reported in the Inventory model
See: https://github.com/italobusi/ietf-network-inventory/issues/2#issuecomment-903086514