[Info] How to handle Adaptive Modulation Data

[kmohr-soprasteria]

Netexplorer needs QAM data.

• There are lots of different qam-states. Counters, measuring the time in a certain modulation are using the transmission-mode-name to denote the modulation scheme
• this name can be "anything" (e.g. something like "1000-16QAM-381808/415520-1" as seen in a RTX380AX device, or "14004" as seen in an AGS20 device)
• there is no common naming
• however, the transmission-modes-list contains a record for each possible modulation scheme which provides the attributes "modulation-scheme" and "modulation-scheme-name-at-lct" which can provide some mapping.
• but there's still a problem: there can be more than one records in the list with the same lct name
  • e.g. seen for Lab AGS20 device 513250006: there are 4 (!) different records for 16QAMStrong

Sample filtered for just 3 attributes:

{
    "air-interface-2-0:air-interface-pac": {
        "air-interface-capability": {
            "transmission-mode-list": [
                {
                    "transmission-mode-name": "28003",
                    "modulation-scheme": 16,
                    "modulation-scheme-name-at-lct": "16QAMStrong"
                },
                {
                    "transmission-mode-name": "14003",
                    "modulation-scheme": 16,
                    "modulation-scheme-name-at-lct": "16QAMStrong"
                },
                {
                    "transmission-mode-name": "56003",
                    "modulation-scheme": 16,
                    "modulation-scheme-name-at-lct": "16QAMStrong"
                },
                {
                    "transmission-mode-name": "112003",
                    "modulation-scheme": 16,
                    "modulation-scheme-name-at-lct": "16QAMStrong"
                },
              ...
            ]
        }
    }
}

[kmohr-soprasteria]

For the example above other attributes in of the transmission-mode-list records are different between the four 16QAMStrong records, e.g. the channel-bandwidth, or am-downshift-level...

                {
                    "transmission-mode-name": "28003",
                    "symbol-rate-reduction-factor": 1,
                    "channel-bandwidth": 28000,
                    "xpic-is-avail": true,
                    "supported-as-fixed-configuration": false,
                    "tx-power-min": -6,
                    "am-downshift-level": -76,
                    "modulation-scheme": 16,
                    "code-rate": 75,
                    "am-upshift-level": -74,
                    "rx-threshold": -80,
                    "modulation-scheme-name-at-lct": "16QAMStrong",
                    "transmission-mode-rank": 28003,
                    "tx-power-max": 21
                },

                {
                    "transmission-mode-name": "14003",
                    "symbol-rate-reduction-factor": 1,
                    "channel-bandwidth": 14000,
                    "xpic-is-avail": true,
                    "supported-as-fixed-configuration": false,
                    "tx-power-min": -6,
                    "am-downshift-level": -79,
                    "modulation-scheme": 16,
                    "code-rate": 74,
                    "am-upshift-level": -77,
                    "rx-threshold": -83,
                    "modulation-scheme-name-at-lct": "16QAMStrong",
                    "transmission-mode-rank": 14003,
                    "tx-power-max": 21
                },

[kmohr-soprasteria]

Info Sven:

das sieht so aus, als hätten sie für jede Bandbreite (7, 14, 28, 56, 112MHz) eine eigene ID, die Modulation ist dann aber immer „16QAMStrong“, die übertragene Bandbreite ist dann aber entsprechend weniger oder mehr. Für die Anzeige der Modulationsstufe ist das korrekt, man muss nur bei der Berechnung der Kapazität/Auslastung vorsichtig sein, weil man hierzu Modulationsstufe und Bandbreite braucht.

[kmohr-soprasteria]

Originally it was envisioned to deliver the data as csv, but with the latest findings that at least makes no sense for the QAM data. (For others like level data or G826 csv would still be fine.)