WP8.1 - GTS -> WIS2 Gateway

[efucile]

26 June 2023

DECISION

• GTS headers shall be used in the TH
• a proposal TH for GTS data on the GC is : cache/a/wis2/xxg/gts_to_wis2_dwd/TTAAii
• DWS Global Cache provides the gateway
• JMA will have internal discussion for decision
• wis2-guide/GTS2WIS2-Gateway.adoc at transition · wmo-im/wis2-guide · GitHub

[Haddouch101]

• DWD GC provides the gateway from GTS to WIS2
• The other GCs shall provide the same gateway

[Haddouch101]

ET-W2AT 15-05-2023:

• GTS headers shall be used in the TH
• a proposal TH for GTS data on the GC is : cache/a/wis2/xxg/gts_to_wis2_dwd/TTAAii

[golfvert]

Let's review this in an upcoming W2AT meeting before closing it.

[josusky]

BTW, there is another proposal to solve the same problem: add property for TTAAii

[KenRJTD]

Additional (Recommended) data/products are circulated over the GTS.
Should these restricted data be published to WIS2? or only Essential (Core) data be published to WIS2.0?

[golfvert]

Everything that is currently exchanged over the GTS should be made available on WIS2 during the transition by the gateway(s). The principle is not to "loose" any data during the transition.
The question though is how to recognize a "recommended" data on the GTS ?
We probably don't want that the gateway(s) to "make" them core at their level.

[KenRJTD]

Res.40 Additional data (pdf files) are available in the middle of WMO web page. https://community.wmo.int/en/activity-areas/operational-information-service/additional-data-and-products
I know Res.40 has already been replaced with New WMO data policy. I haven't seen any list of Recommended data so far. I think we will recognize recommended data by WIS2 metadata in the near future.
My question is, as of today, don't Gateways need to care for the Res.40 additional data?

[antje-s]

A first solution was included in GTStoWIS2 TableA
https://github.com/wmo-im/GTStoWIS2/blob/main/GTStoWIS2/TableA.json
but this is no longer included after the changeover to TTAAii. Perhaps we could differentiate in the pub script itself depending on the TTAAii. But for this a list of all TTAAii headers with recommended data would be useful...as a first solution, we could orientate on the TableA and collect headers including "recomended". It should be ok to try it out and later the list could be completed/corrected.

[antje-s]

We updated topics to
cache/a/wis2/xxg/gts_to_wis2_dwd/data/core/TTAAii/CCCC
and
cache/a/wis2/xxg/gts_to_wis2_dwd/data/recommended/TTAAii/CCCC

Attached the initial list of recommended TTAAii
(startswith-check), which should be corrected/adjusted

---

FA
FC
FR
FT
ISV
ISW
IUA
IUO
JSA
JSP
JSS
JST
JTS
KFA
KFP
KFS
KFT
KSV
KSW
KUA
KUO
L
M
SA
SP
UA
UB
UD
UR
UT
V
Y
Z

[kaiwirt]

After discussion with @golfvert and @6a6d74 we decided that the gateways should publish under the origin TH and the data should be republished by the caches under the cache TH (i.e. GC JMA republishes DWD gateway gts data and DWD GC republishes JMA gateway gts data).

[kaiwirt]

The data id for the WNM from the GTS -> WIS2 Gateways should be unique for each gateway. That is JMA and DWD will choose different data ids for example UUID reflecting the fact, that GTS bulletins with the same header might be different at the different locations.

Therefore there will be two different GTS->WIS2 datasets (one for each gateway). The end-users receiving the GTS data through WIS2 will need to be able to handle duplicates

[golfvert]

After discussion during ET-W2AT the proposed topic hierarchy: cache/a/wis2/xxg/gts_to_wis2_dwd/data/recommended/T1/T2/A1/A2/ii/CCCC
cache/a/wis2/xxg/gts_to_wis2_dwd/data/core/T1/T2/A1/A2/ii/CCCC

[kaiwirt]

Regarding wmo-im/wis2-topic-hierarchy#58 what is suggested to use for the Gateway:

• xxg-gts_to_wis2_dwd
• de-gts_to_wis2
• de-dwd-gts_to_wis2

others?

[golfvert]

We are abandoning the xxa-g stuff. So, not the first one.
Then, recognizing the role of dwd (and jma) I'd go for de-dwd-gts_to_wis2 or even de-dwd-gts-to-wis2
I wonder if we should restric the allowed characters in centre-id to [a-z],[0-9],-
And nothing else.