[FEATURE] Add a Rate function to Register UE in specified Rate

[sambitcom]

Is your feature request related to a problem? Please describe. I am trying to run ./packetrusher multi-ue -n 20000 -l. But the Signaling TP and pps is very low. I want to add a function to support rate of registration so that it runs on that speed. For example 100 Reg/sec .

Describe the solution you'd like A clear and concise description of what you want to happen. If applicable, inform the specifications and sections the defines the behavior.

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

Additional context Add any other context or screenshots about the feature request here.

[linouxis9]

Hi @sambitcom,

Have you tried something like ./packetrusher multi-ue -n 20000 --timeBetweenRegistration 50 --loop. You can reduce the 50 to very low values to increase the signalling rate. Otherwise, we are also working on #31 which would add that kind of parameters!

Thanks and cheers, Valentin

[sambitcom]

Hi Valentin,

I have tried such option but I am getting below error NFO[0051] [GNB] Initiating PDU Session Resource Setup Response FATA[0051] [GNB][AMF] Error sending Initial Context Setup Response: Error sending NGAP message %!(EXTRA syscall.Errno=no buffer space available)

Is there anylogs you required to debug or any workaround available to continue my test?

Regards, SambitM

[linouxis9]

Hi,

It seems that are you sending more messages than your AMF can handle from your gNodeB at once. Starts with --timeBetweenRegistration 500 and reduce the number until just before you get this issue. (@Raguideau in #121, for instance, without a global scenarioExecuter, here it would be hard to limit the speed of messages sent from all UEs to an AMF when we are observing that we are reaching the limit of messages being sent to the AMF)

Thanks, Valentin

[sambitcom]

I have successfully tested with higher Control plane TP above use case by tuning sctp perimeter in linux. You may close this ticket.

[linouxis9]

Hi @sambitcom,

Would you mind sharing the SCTP parameters you tweaked?

Thanks a lot for confirming it worked for you!

Valentin

[sambitcom]

Yes. Sure echo 50000000 > /proc/sys/net/core/wmem_max echo 50000000 > /proc/sys/net/core/rmem_max echo 50000000 > /proc/sys/net/core/wmem_default echo 50000000 > /proc/sys/net/core/rmem_default echo 20000000 > /proc/sys/net/core/netdev_max_backlog

This has increased the buffer in SCTP Association in system Previous ASSOC SOCK STY SST ST HBKT ASSOC-ID TX_QUEUE RX_QUEUE UID INODE LPORT RPORT LADDRS <-> RADDRS HBINT INS OUTS MAXRT T1X T2X RTXC wmema wmemq sndbuf rcvbuf ffff95cc31a99000 ffff95cbdd8af380 2 1 3 0 18 0 8678 1000 1797954 38412 9487 10.60.1.189 <-> *10.60.1.26 7500 2 2 4 0 0 0 1 0 212992 425984

Latest after change AT Core SIDE ASSOC SOCK STY SST ST HBKT ASSOC-ID TX_QUEUE RX_QUEUE UID INODE LPORT RPORT LADDRS <-> RADDRS HBINT INS OUTS MAXRT T1X T2X RTXC wmema wmemq sndbuf rcvbuf ffff95cc33deb000 ffff95cbdd8ab180 2 1 3 0 21 0 23672 1000 2755232 38412 9487 10.60.1.189 <-> 10.60.1.26 7500 2 2 4 0 0 0 1 0 50000000 524288 RAN SIDE SSOC SOCK STY SST ST HBKT ASSOC-ID TX_QUEUE RX_QUEUE UID INODE LPORT RPORT LADDRS <-> RADDRS HBINT INS OUTS MAXRT T1X T2X RTXC wmema wmemq sndbuf rcvbuf ffff8f4cb2fdf000 ffff8f4c72db6880 2 1 3 0 57 1581056 0 1000 199297 9487 38412 10.60.1.26 <-> 10.60.1.189 7500 2 2 10 0 0 3068 1581057 1581056 50000000 50000000