Multi-hop relaying IAB network simulation

[DKhaoyu]

Hello,
I'm trying to simulate the multihop relaying operations. It is mentioned in the bottom of Page. 6 of the paper "End-to-End Simulation of Integrated Access and Backhaul at mmWaves".

In the class "mmwaveHelper.h", I find functions "AttachIabToClosestEnb", "AttachIabToSelectedClosestEnb" and "AttachIabToClosestWiredEnb" to attach the IAB net devices to the closest Enb net device, but not to a IAB net device. So could you please guide me how to realize the multi-hop relaying in simulation? Thanks a lot!

[pagmatt]

Hi Hiroshi, despite the name AttachIabToClosestEnb actually attaches the specified devices to the closest eNB, which is possibly a non-wired one, i.e., an IAB-node. For instance, by calling:

mmwaveHelper->AttachIabToClosestEnb (iabmmWaveDevs, enbmmWaveDevs);

each IAB device in iabmmWaveDevs will connect to either one of the enbmmWaveDevs, or one of the other iabmmWaveDevs.

[DKhaoyu]

Thanks a lot for your helpful feedback!

To test the function of multihop relaying, I modifed the the code in mmwave-iab-grid.cc. For simlicity, I set uint32_t numRelays=2 to enable a 2 hop netowork, by calling

mmwaveHelper->AttachIabToClosestEnb(NetDeviceContainer(iabmmWaveDevs.Get(0)),enbmmWaveDevs); 
mmwaveHelper->AttachIabToClosestEnb(NetDeviceContainer(iabmmWaveDevs.Get(1)),NetDeviceContainer(iabmmWaveDevs.Get(0)));

Unfortunately, the simulator throws an expception in line 1120 of mmwave-3gpp-channel.cc, because the member m_longTerm of channelParams is missing, i.e., the size of channelParams->m_longTerm is 0. To solve the problem, I move the statement CalLongTerm(channelParams) in line 1106 of mmwave-3gpp-channel.cc to line 1119, then the simulator could run without exception. So I wonder whether it is appropriate to do the modification here or did I miss something in the simulation?

[pagmatt]

By having a quick look at it I think that yes, it should work, it may just slow down the simulations a bit. However, I honestly do not have the time to test this out (this is an unmaintained repository, even though we should release an updated version in the next few months/half an year).
My advice is to just try to run the simulation with a different seed for the RNG (by passing -RngRun=N when running the script) . Usually, errors pop up only for certain values of N. It is not an elegant solution, rather a workaround, but it usually works :)

[DKhaoyu]

Thanks for your advice, I will try with different random seeds :)