Groups keyboard shortcuts have been updated
Dismiss
See shortcuts

Slice Setup for Indoor OTA Lab UEs

18 views
Skip to first unread message

Adithya Anand

unread,
Dec 29, 2024, 4:16:58 PM12/29/24
to Powder Users
Hi all 

I am currently running experiments with network slicing in the Indoor OTA lab. I added the changes to config files of the core network and gNB to create an extra slice. Out of the 4 UEs, 2 are on the default provided slice, and 2 are on the newly created slice. UEs registered (through webgui of open5gs) on the default slice connect fine but those connected to the newly created slice do not connect. Further if I were to change the slice config through webgui of open5gs of a UE on the newly created slice to the default slice, that UE connect immediately. My guess is that I need to modify the slice config information of the UEs, was wondering if anyone can share some information about how to make such a change. I have attached some images that could beIndoor OTA slice UE connection issue, img2.jpgIndoor OTA slice UE connection issue, img1.jpg useful. 


Indoor OTA slice UE connection issue, img3.jpg

Dustin Maas

unread,
Dec 29, 2024, 4:33:52 PM12/29/24
to Powder Users

Hi Adithya,

In the future, please copy in text rather than including images, and include a link to your experiment along with your query. In this particular scenario, it would be useful to attach the configuration files you are using. Have you added the second slice to the gNB configuration as well?

-Dustin

--
You received this message because you are subscribed to the Google Groups "Powder Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to powder-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/powder-users/7ddaf0eb-005b-4d1a-9ce4-002d61534cd8n%40googlegroups.com.

Adithya Anand

unread,
Dec 29, 2024, 7:35:38 PM12/29/24
to Powder Users
noted 


I have added slicing in the gNB config. 
gnb_config_indoor_ota.txt
nssf.yaml.txt
amf.yaml.txt

Dustin Maas

unread,
Dec 29, 2024, 7:39:19 PM12/29/24
to Powder Users
Thanks. I’m going to log in and try to attach the UEs while I watch the logs in realtime. 

-Dustin

Dustin Maas

unread,
Dec 29, 2024, 9:15:58 PM12/29/24
to Powder Users

Hi Adithya,

I think the line you added in your nssf pointed to a non-existent nrf, so the nssf couldn’t use it to find the smf for that slice. I just pointed the second slice at the already configured nrf. I also updated the credentials for ota-nuc{3,4} to use sst:2 sd:2 (all of the UEs were still configured to use sst:1 sd:1 when I logged in to check things out). Now I know just asked you to prefer text over images, but in this case an image is more useful, so see attached and note that the UEs attached to ota-nuc{3,4} are now using the desired second slice.

-Dustin

Screenshot 2024-12-29 at 17.50.13.png

Adithya Anand

unread,
Dec 30, 2024, 1:47:15 PM12/30/24
to Powder Users
Thanks Dustin 

I was able to test the system last night and could see it working, I forgot to mention before but I was using a modified srsRAN project code but the system still worked. 

Last night I overlooked checking the config files for open5gs. From my understanding of what you said, did you change the addr of the new slice in nssf to the one already provided, so 127.0.0.10? If I had wanted to make a new addr for my slice should I make a new entry in nrf with that addr? Does have a new addr in nssf change the behavior of slices? Would it be possible to share the config files you ended up modifying for the core. 

yes I figured I had an issue with UEs being configured to the base slice and not the new one. Could you share the commands you used to reconfigure the slice information for the UEs. 

Best
Adithya Anand

Dustin Maas

unread,
Dec 30, 2024, 2:19:41 PM12/30/24
to Powder Users

Hi Adithya… answers inline.

On 30 Dec 2024, at 10:47, Adithya Anand wrote:

Thanks Dustin

I was able to test the system last night and could see it working, I forgot
to mention before but I was using a modified srsRAN project code but the
system still worked.

I tested with the modified version and gnb config that you had been using, in order to diagnose the problem.

Last night I overlooked checking the config files for open5gs. From my
understanding of what you said, did you change the addr of the new slice in
nssf to the one already provided, so 127.0.0.10?

Yep. The only thing I changed was that line in /etc/open5gs/open5gs-nssfd.service, since no NRF existed at 127.0.0.19, and it wasn’t clear why that was set that way.

If I had wanted to make a
new addr for my slice should I make a new entry in nrf with that addr?

My understanding is that the line you changed was the address for the NRF to be used for the second slice, which I changed to be the same as the NRF that is used for the first slice, since there was clearly not a second NRF running.

Does have a new addr in nssf change the behavior of slices?

Just having a second NRF running would not change the behavior of the slices.

Would it be possible to share the config files you ended up modifying for the core.

The only change I made to the config files was the address change in the nssf.

yes I figured I had an issue with UEs being configured to the base slice
and not the new one. Could you share the commands you used to reconfigure
the slice information for the UEs.

I did not change anything at the UEs themselves. I simply used the Open5GS Web UI to update the SST and SD for two of the UEs.

Best
Adithya Anand

-Dustin

On Sunday, December 29, 2024 at 8:15:58 PM UTC-6 Dustin Maas wrote:

Hi Adithya,

I think the line you added in your nssf pointed to a non-existent nrf, so
the nssf couldn’t use it to find the smf for that slice. I just pointed the
second slice at the already configured nrf. I also updated the credentials
for ota-nuc{3,4} to use sst:2 sd:2 (all of the UEs were still configured to
use sst:1 sd:1 when I logged in to check things out). Now I know just asked
you to prefer text over images, but in this case an image is more useful,
so see attached and note that the UEs attached to ota-nuc{3,4} are now
using the desired second slice.

-Dustin

On 29 Dec 2024, at 16:38, Dustin Maas wrote:

You received this message because you are subscribed to the Google Groups
"Powder Users" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to powder-users...@googlegroups.com.
To view this discussion visit

--
You received this message because you are subscribed to the Google Groups "Powder Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to powder-users...@googlegroups.com.

Reply all
Reply to author
Forward
0 new messages