Throughput in multivendor environment nokia & huawei

[oswe]

Hi guys we are facing comparison between those 2 vendors . formulas are the official ones but nokia stays behind. Cant reach the throughput that huawei provides. Anyone faced this comparison of throughput? We suspect transmission issue in the case of nokia.

[electron]

Hi guys we are facing comparison between those 2 vendors . formulas are the official ones but nokia stays behind. Cant reach the throughput that huawei provides. Anyone faced this comparison of throughput? We suspect transmission issue in the case of nokia.

Hi,

Assuming there is no network configuration issue or transport related bottleneck and only KPI formula differences. I would like to highlight that in Huawei they usually use full buffer UE into consideration that's why statistically they show higher throughput and excluding small samples . That might be a case


B
R

[oswe]

That is very difficult to prove we are thinking exactly the same. If any would have this proof in hands it would be very welcomed thanks electron

[ramboswrath]

Hi guys we are facing comparison between those 2 vendors . formulas are the official ones but nokia stays behind. Cant reach the throughput that huawei provides. Anyone faced this comparison of throughput? We suspect transmission issue in the case of nokia.

The scenario is kinda vague. are you comparing stats / swap comparison or different sites?

[general.1979]

Hi guys we are facing comparison between those 2 vendors . formulas are the official ones but nokia stays behind. Cant reach the throughput that huawei provides. Anyone faced this comparison of throughput? We suspect transmission issue in the case of nokia.

turn off huawei tool

[electron]

That is very difficult to prove we are thinking exactly the same. If any would have this proof in hands it would be very welcomed thanks electron

the counter explanation in E******* and Huaweii gives you the hint


B
R

[oswe]

The scenario is kinda vague. are you comparing stats / swap comparison or different sites?

scenario is kinda vague if you are considering low experienced networks.
we have eliminated many variables in drive tests, city morfology, existance of 1800 and 2600 networks... many variables were checked and in the end Nokia is offering a worst performance than huawei.
comparison is simple as Ive mentioned... no need for scenario as live network is NOT a lab...
simple put throughput vs users is getting rapidly worst in Nokia than huawei many countries have observed that...

[jan74]

scenario is kinda vague if you are considering low experienced networks.
we have eliminated many variables in drive tests, city morfology, existance of 1800 and 2600 networks... many variables were checked and in the end Nokia is offering a worst performance than huawei.
comparison is simple as Ive mentioned... no need for scenario as live network is NOT a lab...
simple put throughput vs users is getting rapidly worst in Nokia than huawei many countries have observed that...

And how is this being measured? Via KPI or Speedtest? If its KPI, I would not be too worried as there are so many variables. If its Speedtest, assuming sample size is large and comparable, then there could be an issue. But you would need to dig deep into the test logs and check MCS allocations, CQI, mimo etc. Was the same handset used? There are so many variables so its difficult to know without all of the details.

[oswe]

And how is this being measured? Via KPI or Speedtest? If its KPI, I would not be too worried as there are so many variables. If its Speedtest, assuming sample size is large and comparable, then there could be an issue. But you would need to dig deep into the test logs and check MCS allocations, CQI, mimo etc. Was the same handset used? There are so many variables so its difficult to know without all of the details.

yeah i agree but customer in Brazil is comparing both vendors ... difficult to discuss...
we suspect the formula in huawei is not considering full buffered tti's , or discarding empty ones...
nokia offers 2 throughput formulas and if you consider the formula discarding ttis where you US was not scheduled, throughput is VERY similar to huawei... (with similar number of users, of course)
so ok, this is dealt in many countries by colleagues in India , europe , etc... i just wanted to share in case anyone is having this headache. ideal would be for us to prove huaweis formula counting or not counting some tti´s i let you know if we reach that point. thanks guys