Mesh message queuing

[AmiguelS]

Good afternoon,

I have been doing some tests with the nRF SDK for Mesh v1.0.0, and have a question regarding it.

I'm using a fw based of the Light Switch example, to try to see how many messages and how fast it's possible to queue without loosing any of them.

What I noticed, however, is that the server is able to send messages a lot faster than the client. What parameters could affect this? Are they configurable?

Could it be related to the use of RTT on the client?

The same exact hardware is being used for both client and server.

Thanks.

[tesc]

Hi,

For the light switch example, the client uses reliable messages and the server does not. Non-reliable messages are a lot faster than reliable messages, and so you should expect a significant difference between the two.

There should be no difference if you use the same method in both ends for sending the same message.

Regards,
Terje

[AmiguelS]

Hello, Thanks for the reply. Where can I configure the delivery method (reliable/non reliable)?

[AmiguelS]

To convert the example from reliable messages to non-reliable messages, is it only needed to use ``access_model_publish(...)`` instead of ``access_model_reliable_publish(...)`` ?

[tesc]

Yes. Those are the API calls for non-reliable and reliable messages respectively.

Please note that only for the reliable publish you will get a response acknowledging that the message is received. In order not to loose anything you would have to use reliable publish.

There is a trade-off here, between reliable (loss-less) communication and speed.

[Pooja]

I am using  ``access_model_reliable_publish(...)``  but it gives the error?

it return status value as 14.

<t:      57159>, nrf_mesh_sdk.c,   72, APP_ERROR: ..\nrf5_SDK_for_Mesh_v1.0.1_src\examples\light_switch\client\src\main.c:345 code 14

What's wrong with this?