Simulation experiment "SimpleRelayCongestion" not working

[kaleal]

Hi guys,

I am having problems running SimpleRelayCongestion experiment. I just ran the experiment without any modification. The error occur when using the RINASim VM and also with a fresh installation on Windows and Linux OS. Screenshot and module output log (partial) is attached.

During all experiment the host2 doesn't joing the Layer11 DIF. I think this is the root cause of the problem. There are several RTTEstimator errors reported, but I think is just because the host2/ipcProcess1 is not joined into Layer11 DIF.

module_log.txt

[gaixas1]

Hi. Yes, most probably that's the problem. Add this line **.isSelfEnrolled = true somewhere at the begining of the ini file.

2016-07-30 21:18 GMT+02:00 Kleber Leal notifications@github.com:

Hi guys,

I am having problems running SimpleRelayCongestion experiment. I just ran the experiment without any modification. The error occur when using the RINASim VM and also with a fresh installation on Windows and Linux OS. Screenshot and module output log (partial) is attached.

During all experiment the host2 doesn't joing the Layer11 DIF. I think this is the root cause of the problem. There are several RTTEstimator errors reported, but I think is just because the host2/ipcProcess1 is not joined into Layer11 DIF.

[image: simplerelaycongestion_error] https://cloud.githubusercontent.com/assets/15032576/17272540/8ba31b3a-566f-11e6-86bc-bbe22f352917.png module_log.txt https://github.com/kvetak/RINA/files/392055/module_log.txt

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/kvetak/RINA/issues/15, or mute the thread https://github.com/notifications/unsubscribe-auth/ABY86In5I2FmpPBVt7H_jeobHSNqazIiks5qa6OEgaJpZM4JY39Y .

[kaleal]

Thank you for replying, @gaixas1.

I made it, but unfortunately the problem remains. Adding .isSelfEnrolled = true the simulation aborted even earlier (at 5.006s), then a set host2.ipcProcess1.**.isSelfEnrolled = true. After that, the ipcProcess1 is checked as enrolled, but the simulation running aborted as 55s as it done previously.

Even ipcProcess1 enrollment is checked, the RTTEstimator erros messages remains.

Kleber

[gaixas1]

Try running it un debug to see where it crashes. I'm not on my desktop so cannot do more than sugestions on this.

El 30/7/2016 22:27, "Kleber Leal" notifications@github.com escribió:

Thank you for replying, @gaixas1 https://github.com/gaixas1.

I made it, but unfortunately the problem remains. Adding _.isSelfEnrolled = true the simulation aborted even earlier (at 5.006s), then a set _host2.ipcProcess1..isSelfEnrolled = true. After that, the ipcProcess1 is checked as enrolled, but the simulation running aborted as 55s as it done previously.

Even ipcProcess1 enrollment is checked, the RTTEstimator erros messages remains.

Kleber

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/kvetak/RINA/issues/15#issuecomment-236388241, or mute the thread https://github.com/notifications/unsubscribe-auth/ABY86JZ7Hif-U2jWoqW1Bu7BI9zl9vzLks5qa7OngaJpZM4JY39Y .

[kaleal]

I wasn't able to find a solution yeat. Can anyone help me? I cannot run this experiment.

Kleber

[kvetak]

Okay, let me help you. I will investigate your problem tomorrow and write you my thoughts.

[kaleal]

Thank you.

Kleber

[kvetak]

Multiple things were wrong with this scenario. I have addressed them in the following commit: https://github.com/kvetak/RINA/commit/380a46c6a5bf83f25e40ca6eb5504fcd142adde7

Among problems were: 1) Improper usage of flow preallocation in config.xml. This technique is deprecated since Enrollment. 2) Improper usage of AE name, where App2 was used (thanks to config.xml) as synonym AppErr which always rejects flow allocation. 3) The most important was timing issue. Link between interiorRouter and host2 has delay of 5 seconds for transmission (see SimpleRelayCongestion.ned). Hence, host1->host2 ping had started too early before even host2 was properly enrolled to Layer11.

Remember that Enrollment is integral part of RINA communication. Without Enrollment (at least attempt) there are no management flows, which are mandatory for IPC management messages. Therefore, using global **.isSelfEnrolled = true is extremly undeterministic technique and I urge everyone not to do this!

Scenario is currently "repaired" in the following way:

• preenrollment of host2 works
• data flow allocation works
• ping transmission occurs

I cannot determine the state of congestion since it is not my branch of RINA expertise. Hence, I kindly ask my colleague to continue...

@screw > Please, test the congestion (check EFCP parameters and their relevance to current omnetpp.ini settings) and provide feedback. If you consider this scenario is problematic, we will move it to playground folder.

[Kleberleal]

Thank you, Vladimir. It's working properly right now.

Kleber