How to install two LMs in one ETS project and use them for different purposes

[Takayuki Miura]

Hello!

Regarding the subject, I would like to set up two LMs in one ETS project and use them for different purposes.

LM① = Only for sending commands
LM② = Only for receiving status feedback

I would like to use them for the above purposes.
It is like the image in the attached figure.

I have also attached the backup files of LM① and LM② created for testing, as well as the ETS project.
Due to capacity restrictions, I have created a download link for the ETS project.
Please download it from the link below.

https://firestorage.jp/download/c39910c8...a4877da3cf

I have a few questions.

1. When importing one ETS project into LM, is it possible to filter the group address to be imported?
→As a test, I created an environment to use two LMs in one project.
I have also attached a backup of the LM and the ETS project, so please check them.

2. For example, I would like to know how to use the [Event script] set for the group address in LM① to write a value to the group address in LM② or update the numerical value.
→This time, in the test environment, I created it to try a process such as [writing true to 20/0/1 in LM① in the event script written to 21/0/1 in LM②].
If there is a better way, please let me know.

3. I came up with the idea of using two LMs for different purposes in order to distribute the load on one LM, but have there been any cases where this type of operation has actually been used?
Also, when using two LMs for one project, what is the general purpose for which they are used?

I would appreciate your reply.

[Daniel]

Why would you want to do this? Connecting two LMs to the same TP makes zero load limitation as anyway all telegrams will go to both LMs.
If you have too big traffic you should divide the lines and connect one LM to one ans second to second and then they will talk to each other via IP.

[Takayuki Miura]

Sorry. I didn't explain it well enough.

I'm assuming that the LM and the external cloud server will use API to send and receive commands, status, etc. continuously without interruption.

LM① = Dedicated to receiving commands from external servers

LM② = Dedicated to sending status to external servers

I would like to use it like this.

Currently, one LM is sending and receiving APIs for a hotel project with about 1050 group addresses, but packet loss can occur when performing a process such as turning all facilities on at check-in.

Rather than being a problem of TP load, it is often the case that the LM load is quite high due to the frequent use of API to send and receive data.

Therefore, I would like to use different LMs for sending and receiving APIs for the same TP.

This project is a building about 1.5 times the size of the hotel, so it is assumed that there will be about 1600 group addresses.

[Daniel]

I still think your problem is on the TP side. Your proposal makes no sense even for your application. If you want you can connect second LM just on IP side and disconnect from TP and use it as the API communication. LM to LM will talk over KNX IP.
What I wrote earlier is still the best solution, to divide this project in smaller ones where one LM will be enough to maintain KNX and your API.
Make sure that your cloud is not writing constantly to KNX as this will kill KNX and LM.

[Takayuki Miura]

Hello.

It's been a while.

As per your advice, I am considering placing KNX devices on separate lines.
I'm thinking of creating it in the format shown in the attached image.

I intend to import only the group addresses related to each line into each LM as a single ETS project.

I would also like to know how to set up the KNX connection in the LM's system config when using it this way.

I would appreciate it if you could tell me if the following usage is physically possible and, if possible, how to do it.

① Write a value to the group address in the LM in Line 2 using the event script in the LM in Line 1

② As mentioned above, how to send and receive data between LMs when the two LMs are connected only by IP

→ I expect this is possible if you use API, MQTT, etc., but I would like to know if there is a good way that is possible only between LMs in the same network.

③ If I want to use only LM1 (placed on Line 1) as the LM used as the API and MQTT interface, how should I design it to monitor the status of the group address on the Line 2 side and send commands from an external server?

Both of these concerns are things I would like to consider in order to achieve my goals of distributing the number of KNX TP telegrams by dividing the lines with LMs, and, if possible, limiting the number of LMs that send and receive data with the outside world using MQTT etc. to just one.

I would appreciate any advice you could give me.

[admin]

With the default configuration LM acts a KNX/IP router meaning all telegrams are exchanged transparently between KNX/TP and KNX/IP. You can allow or block only certain group addresses between TP and IP if needed.

If you want to get all KNX traffic into a single LM then you either need 3 LMs or 1 LM + 2 IP Routers. In both cases the main LM won't use a KNX/TP connection, only KNX/IP.