Getting different results in express mode and hybrid mode (run with full animation, then express)

[S.Hedayat Hosseini]

Dear All,
We use OMNet++ 5.2 and INET 3.6.3.
We defined a centralized module to generate traffic and process statistics. When a host received a packet, it passes (by function call, not sending a message) some information to the centralized module which collects information of all hosts.
Unfortunately, when we run network in hybrid mode (for example, first, we run network in "full animation" mode and after a while, we choose express mode to continue), we get a different result from the complete express mode(When we completely run the simulation in express mode).
In a test, the results of the second state were completely wrong; for example, even the number of received messages was greater than the number of sent messages.
Has anyone experienced this problem?
Thank you.

[Levente Mészáros]

There's a known issue in INET 3.6.x with some mobility modules that affects the results when viewing the animation in the runtime user interface. Whenever the user interface refreshes the screen it needs to query the positions from the mobility modules. Unfortunately, the mobility trajectory computed by several current mobility models depend on when the positions are queried. The difference is infinitesimal though, it is caused by interpolating from the last known position as opposed to some previous stable state. This wasn't an issue before OMNeT++ introduced animation in the runtime user interface. This is still not an issue unless you are viewing the animation, and it still should not affect your statistical results due to the difference being so small. The upcoming INET 4.0 will fix this non-determinism in mobility models. I'm not sure if this is the reason for the difference you are experiencing but it may be.

Best regards,

levy

On Wed, Feb 14, 2018 at 10:18 PM, S.Hedayat Hosseini <niava...@gmail.com> wrote:

Dear All,
We use OMNet++ 5.2 and INET 3.6.3.

We defined a centralized module to generate traffic and process statistics. When a host received a packet, it passes some information to the centralized module which collects information of all hosts.

Unfortunately, when we run network in hybrid mode (for example, first, we run network in "full animation" mode and after a while, we choose express mode to continue), we get a different result from the complete express mode(When we completely run the simulation in express mode).
In a test, the results of the second state were completely wrong; for example, even the number of received messages was greater than the number of sent messages.
Has anyone experienced this problem?
Thank you.

--
You received this message because you are subscribed to the Google Groups "OMNeT++ Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to omnetpp+unsubscribe@googlegroups.com.
Visit this group at https://groups.google.com/group/omnetpp.
For more options, visit https://groups.google.com/d/optout.

[Hedayat Hosseini]

Dear levy,
Thanks for your response.
Regarding statistics, we use StationaryMobility and collect statistics from our app layer module. Since we did not make changes to the UDP module, we investigated statistics accumulated by UDP module. Unfortunately, this problem also exists in the statistics of the UDP module. Results (for 60.2s) can be seen in attached figures. Difference of delivery ratio is about 30% in two experience.

Thank you very much!

Best,
Hedayat

در پنجشنبه 15 فوریهٔ 2018، ساعت 12:05:31 (UTC+3:30)، levy نوشته:

To unsubscribe from this group and stop receiving emails from it, send an email to omnetpp+u...@googlegroups.com.

[Levente Mészáros]

StationaryMobility is clearly not the reason. It's hard to say anything about what is wrong without being able to reproduce the problem. You could try to narrow it down and send a bug report with detailed steps to reproduce.

Best regards,

levy

On Fri, Feb 16, 2018 at 11:51 AM, Hedayat Hosseini <niava...@gmail.com> wrote:

Dear levy,
Thanks for your response.

Regarding statistics, we use StationaryMobility and collect statistics from our app layer module. Since we did not make changes to the UDP module, We investigate statistics accumulated by UDP module. Unfortunately, this problem also exists in the statistics of the UDP module. Results (for 60.2s) can be seen in attached figures. Difference of delivery ratio is about 30% in two experience.

To unsubscribe from this group and stop receiving emails from it, send an email to omnetpp+unsubscribe@googlegroups.com.