Let’s imagine we have a VIP area were it is supposed to be mass event (kind of exhibitions).
In that area we have indoor and outdoor cells. 2-3-4G networks.
There is only info that about 2000 people will come to that event.
No Volte. only CS voice.
Let’s assume that all users support 4G.
I need to estimate capacity for 4G network in BH.
There are 46 4G cells (25 cell with 10MHz and 21 cells with 15MHz).
Of course it is obvious that not all 2000 people will use 4G network simultaneously.
Do you have any similar experience and considerations about such tasks?
Is it possible to tell number of 4G users in Busy Hour: 5% or maybe 15% is it enough?
Any ideas would be grate to hear!
In mass events as mentioned scenario mostly traffic is for Upload, which requires high PUCCH capacity.
Solution like massive MIMO can help here or team can plan to increase PUCCH bandwidth for CQI to cater more RRC users and increase max num RRC license to max.
Feature like idle mode load balancing and inter Freq load balancing can help further.
Parallel uplink throughput improvement feature like UL Comp can help.
Users having MO Call can move to lower rat direct by SBHO feature so PRB utilisation can be optimum.
Every operator has their own way to determine QoS, hence it yields on empiric values for number of guaranteed user per cell.
In my case,
FDD 20Mhz : 185
TDD 20Mhz : 185
FDD 15Mhz : 160
And so on.
The “guaranteed user” number is subjective to each operator strategic point of view, most of my case: user throughput + latency based on historical OSS measurements.
If there are already 46 cells there should be no issue. At events, uplink capacity is challenged. And usually, during break time, number of active users can be anything between 30% to 50% from those 2000 users.
This is because user traffic is voice only (network throughput capacity requirement is only i.e.,8kbps, no need CA, but need carefully schedule multiuser in each frame to handle average 50 users per cell). Total capacity requirement is not high, for example 2000*8kbps=16Mbps for the UL, considering network overhead, radio overhead and efficiency 50% and 50%, the PHY rate is about 64Mbps for UL. as @stev.aryanto mentioned, it can be easily handled with the total 565MHz, which is time-shared by both DL and UL (so we provide budget of 128Mbps for UL). We need schedule around 6 users per slot and accommodate around 50 users each 8 slots to fully utilize 8 the HARQ processes. In each slot, 6 users user OFDMA sharing 10MHz or 15MHz bandwidth, each one only have around 1-2MHz bandwidth. If MCS rate is >1 bps/Hz, each one can get 1-2Mbps throughput. That’s enough.
