Statistics 'macCellThroughput(D2D|Ul|Dl)' use static member for total byte count producing overflow and wrong data in multi basestation scenarios

The statistics are defined in LteMac.ned

        @signal[macCellThroughputUl];
        @statistic[macCellThroughputUl](title="Cell Throughput at the MAC layer UL"; unit="Bps"; source="macCellThroughputUl"; record=mean);
        @signal[macCellThroughputDl];
        @statistic[macCellThroughputDl](title="Cell Throughput at the MAC layer DL"; unit="Bps"; source="macCellThroughputDl"; record=mean);
        @signal[macCellThroughputD2D];
        @statistic[macCellThroughputD2D](title="Cell Throughput at the MAC layer D2D"; unit="Bps"; source="macCellThroughputD2D"; record=mean); 

The recording of the statistics takes place the extractCorrectPdus method of the respective HarqBuffer ( normal or
d2d enabled case). The cell throughput is calculated by dividing the total amount of bytes received by the total amount of elapsed time.

The problem is that totalCellRcvdBytes_ is a static unsigned int. See LteHarqBufferRx .

• This means if multiple eNB's/gNB's are simulated, they will add up their received total.
• The variable can only hold 2^32 which is easily read when all eNB's/gNB's write to the same variable.

Possible fix.

Use INET ThroughputFilter result filter implementation.
See for instance the ActivePacketSourceBase module of the INET queuing API. Here the throughput is calculated
by the ThroughputFilter using statistics and result filter tools provided by OMNeT. It also allows to configure the
interval length using ini files on a simulation by simulation basis.

@statistic[dataRate](title="data rate"; source=throughput(packetPushed); record=vector; unit=bps; interpolationmode=linear);

Start with LteHarqBufferRx.h and the extractCorrectPdus methods in LteHarqBufferRx.cc and LteHarqBufferRxD2D.cc, then compare the referenced INET ThroughputFilter and ActivePacketSourceBase implementations. Confirm how the current static byte counter affects multiple base stations and overflow; done means throughput is correctly separated and measured without the shared 32-bit limitation.