View Issue Details
ID | Project | Category | View Status | Date Submitted | Last Update |
---|---|---|---|---|---|
0004677 | 10000-012: Discovery | Api Change | public | 2019-03-13 11:19 | 2022-05-10 16:48 |
Reporter | Jouni Aro | Assigned To | Randy Armstrong | ||
Priority | normal | Severity | minor | Reproducibility | have not tried |
Status | closed | Resolution | fixed | ||
Fixed in Version | 1.05.02 RC1 | ||||
Summary | 0004677: Format of Client EndpointUrl for reverse connections | ||||
Description | The specification (1.04, Part 6. 7.1.3) defines: "The Server needs to be configured and enabled by an administrator to connect to one or more The text mixes different EndpointUrls, so it is a bit hard to follow, which is which, but it also talks about 'the Client EndpointUrl' giving an impression that the reverse socket that the client opens, should be referred with a similar EndpointUrl as the server's EndpointUrl, i.e. 'opc.tcp://clienthost:clientport'. I am not sure if it is a good idea to use a similar URL - or if this even is the intention. Anyway, it would be good to clarify this. If the administrator is to configure the EndpointUrl of a GDS, then he also needs to configure the ApplicationUri of the client - or how is the server supposed to find the client from the GDS? And even worse: how can the GDS provide the reverse endpoint of the client? Part 12 does not talk about EndpointUrls for clients... | ||||
Tags | No tags attached. | ||||
Commit Version | |||||
Fix Due Date | |||||
|
Hmm. After a few more readings, I get it that "For each Client, the administrator shall provide an ApplicationUri and an EndpointUrl So, yes, there is an ApplicationUri that could be used to fetch the EndpointUrl of the client - and Part 12 actually mentions that the EndpointUrl should be prepended with 'inv+', if it is defined for the client. But would be good to clarify if this is the format that is expected to be used in server configuration or if it is left totally open. After all, the server needs to configure the hostname and port only, so the format is free in that sense. ('inv+' sounds a bit odd; mainly we are mixing terms 'inverse' and 'reverse' here, which I don't like very much. In that sense, 'rev+' would have been better, but apparently, I am late, once again.) |
|
Text in Part 12 now clarifies what needs to happen: 1) Applications that support Reverse Connect have an ApplicationType of "Client" or ClientAndServer"; Configuration applications/servers looking for endpoints can pass the ApplicationUri into the QueryApplications on a GDS to get the DiscoveryUrls. |
|
Agreed to changes edited in telecon. |
Date Modified | Username | Field | Change |
---|---|---|---|
2019-03-13 11:19 | Jouni Aro | New Issue | |
2019-03-15 08:15 | Jouni Aro | Note Added: 0010055 | |
2019-03-15 08:16 | Jouni Aro | Note Edited: 0010055 | |
2021-04-14 17:35 | Jim Luth | Project | UA => 10000-006: Mappings |
2021-08-19 05:06 | Randy Armstrong | Project | 10000-006: Mappings => 10000-012: Discovery |
2021-11-30 17:48 | Jim Luth | Assigned To | => Randy Armstrong |
2021-11-30 17:48 | Jim Luth | Status | new => assigned |
2022-03-03 08:55 | Randy Armstrong | Status | assigned => resolved |
2022-03-03 08:55 | Randy Armstrong | Resolution | open => fixed |
2022-03-03 08:55 | Randy Armstrong | Note Added: 0016144 | |
2022-05-10 16:48 | Jim Luth | Status | resolved => closed |
2022-05-10 16:48 | Jim Luth | Fixed in Version | => 1.05.02 RC1 |
2022-05-10 16:48 | Jim Luth | Note Added: 0016699 |