RNS reliable delivery #206
-
From what I can tell from the source code, I think only the resource transfer would be considered reliable delivery--it has a retry mechanism. Request/response has receipt acknowledgement but does not have retries. Is this correct? |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 2 replies
-
You are more or less correct in your observations. The low-level Now, the higher-level This behaviour is partially implemented currently. In some cases, the reliability and retry mechanism is automatically inherited from the So the gist of it is:
|
Beta Was this translation helpful? Give feedback.
You are more or less correct in your observations. The low-level
Packet
API is best effort, but you do have the ability to get a per-packet delivery (or failure) notification with the PacketReceipt API. The intention is that it is up to your application how to react in terms of undelivered packets. You can implement retries, or in some cases you might want to just "send and forget", for simple telemetry and status, for example.Now, the higher-level
Request
API is meant to be reliable, as in having automatic retries on failure. This makes sense to have here, since it is a much higher-level function, closer to the application code, and should at least make a pretty good effort to complete …