Difference in Downlink Packet Loss Rate Between LTE and NR Modules in Simulated Scenarios of Base Stations on Satellites

[nile]

Dear 5G-LENA Community Members,  

Against the background of 3GPP NTN (Non-Terrestrial Network) base stations on satellites, I conducted simulations by integrating a satellite mobility model into the base station setup. When using the LTE and NR modules of ns-3 respectively to simulate low Earth orbit (LEO) satellite communication scenarios, I observed a notable phenomenon: when a single user is within the communication range of multiple satellites (base stations, i.e., eNB/gNB), the downlink of the LTE framework exhibits obvious periodic packet loss fluctuations. Moreover, such fluctuations intensify as the number of visible satellites increases. However, under the NR module, with the exact same configurations (including satellite mobility model, custom path loss model, and user position), the downlink performs stably without any packet loss.  

I have ruled out the path loss model as the cause of this issue. I hypothesize that this difference may stem from the inherent mechanism disparities between the two modules in handling multi-base station connections. One possible explanation is that the handover process of the NR module is not yet fully optimized—its user equipment (UE) may not actively perform measurement reporting and handover negotiation with all visible satellites as LTE does, thereby avoiding complex signaling interactions and potential conflicts. Another possibility is that in the LTE framework, downlink transmissions from multiple eNBs to the same user cause co-channel interference, while the beam management or resource scheduling mechanism of NR may have inadvertently mitigated this problem.  

I would like to consult the community on the following questions:  
1. Are there any known reports regarding the downlink instability of the LTE module in multi-satellite (multi-eNB/gNB) scenarios?  
2. Is the current connection management strategy of the NR module indeed simpler than that of LTE, lacking complete neighbor cell detection and handover processes?  

Any insights into the key differences between these two modules in terms of underlying scheduling, interference coordination, or handover logic would be extremely helpful to my research.  

Thank you in advance for your valuable time and contributions.  

Best regards

[Gabriel Ferreira]

>  One possible explanation is that the handover process of the NR module is not yet fully optimized—its user equipment (UE) may not actively perform measurement reporting and handover negotiation with all visible satellites as LTE does, thereby avoiding complex signaling interactions and potential conflicts.

Handover of NR is not tested nor advertised. We are still working on it.
Measurements are currently being reported not as frequently as in LTE. Also working on it.
RLF is currently not used, so you just continue transmitting. In LTE it is properly detected and retries to reestablish the connection.
There are many things missing, and all of them can and will impact these results.

As far as I know, both LTE and NR do not handle propagation delay either. Which is minimal on terrestrial networks, but very relevant in satellite.