How many (maximum number of) Users can be simultaneously active in a 5G cell/gNodeB?

Hello Experts.

I have seen this question here, long time ago: How many (maximum number of) Users can be simultaneously active in a LTE cell/eNodeB?

Now, i would like to rise same question, but for 5G:

How many users can be simultaneously active in a 5G cell?

I know it depends on several factors (bandwidth, hardware…). Is there also a vendor dependency on this value?

So, what would be the maximum number of concurrent (simultaneous) users - practical and theoretical value - that we can have, considering the main possible scenarios/factors?

Probably there will be some “it depends”… in this case, can we have a table, with the average and/or maximum number of 5G users per cell, for different scenarios?

2 Likes

Really good question. I don’t know the answer, but indeed it’s a good table to have, please share if you find.

I read and article long time ago, saying:

“5G is expected to support up to 1 million connected devices per .38 square miles, compared to around 2,000 connected devices per .38 square miles with 4G.” (This is the link for the article, but it is very old - 2017).

Me too. Only have knowledge for LTE (as in this great article: LTE And The Number Of Simultaneously Connected Users – WirelessMoves).

But for 5G, no idea… :frowning:

It is true to have millions of connected users in case benefitting from all functionalities of 5G (URLLC, EMBB, M2M).

You can find thousands of cars in a crowded street and hundreds of thousands of IoT devices or sensors in the same street…let’s here, as 5G global view, you can serve thousands or much more per cell (via different slices).

Coming back to the usage of 5G for BB domain only (similar to 4G or 3G), i think it is limited to SW level per vendor, BW, license capacity as well HW and configuration of the 5G cell (some alarms may affect negatively the 5G cell performances and hence the number of connected users…)

1 Like

Simultanously active or simultanously scheduled in one TTI?

Simultanously active have dependancy on PUCCH resources, hardware limit and hard limit set by setting

Usually, I value i saw so far is 1200 per cell.

Simultanously scheduled in one TTI: max 16 as per 3gpp.
But can be limit through parameters.

3 Likes

I’ve seen a lot of answers, like:

  • depends on cell likes pico/micro/metro cell

  • haha… the magic answer for this question is “depends on traffic model”… :rofl:

  • Thousands or millions depends of the Network Slicing framework which include different services and use case escenarios.

I agree, in part… :roll_eyes:
But if i answer this to my boss, then he will tell me that only one gNB is enough… :woozy_face:

What is the calculation method of Simultaneously Active UEs in gNB?

At the measurement time, whatever UEs have buffer available in RLC, those all are active UEs.

While dimensioning a node a factor is important in capacity part: Simultaneously Active users

How to calculate? Actually I mean that.

Won’t max simultaneous active users depend on the max HARQ processes - which is 16 in 5G?

How do 100 users share one 5G cell?

A 5G cell does not provide every connected user with a permanently reserved portion of its total capacity.

The gNB scheduler continuously distributes available time-frequency radio resources among active users.

This animation begins with a lightly loaded cell.

Only a few users are active, so each user can receive a relatively large share of the available resources. Some resource blocks remain unused because total demand is below cell capacity.

As more users become active, the scheduler must divide the same radio resources among a larger group.

The animation progresses through five illustrative phases:

  1. Light load

A small number of users are active. Radio resources remain available, and each user can receive a larger average share.

  1. Moderate load

More users begin transferring data. The scheduler rotates resource allocations across devices and applications.

  1. Heavy load

Most time-frequency resources are continuously occupied. Average throughput per user begins to decrease.

  1. Congestion

Combined offered demand exceeds the illustrative cell capacity. Packets may wait in queues, application throughput may fall, and latency may increase.

  1. Active QoS management

Priority traffic receives protected scheduling opportunities while lower-priority background traffic receives a smaller or delayed allocation.

The resource grid in the animation represents an illustrative combination of frequency blocks and time slots.

Different colors represent:

• Normal data traffic
• Video traffic
• Priority traffic
• Unused resources

The scheduler’s decisions can depend on several factors:

• Channel quality
• Traffic demand
• QoS requirements
• User priority
• Buffer occupancy
• Retransmissions
• Application type
• Fairness policies
• Available spectrum
• MIMO layers and radio conditions

A user with better radio conditions may transfer more data using the same amount of radio resources because a more efficient modulation and coding scheme may be possible.

A user near the cell edge may require more resources, additional retransmissions, or a more robust coding scheme to deliver the same amount of application data.

The animation uses a simplified model:

Illustrative cell capacity: 1 Gbps
Illustrative demand per active user: 15 Mbps

With 10 active users, combined demand is approximately 150 Mbps.

With 60 active users, demand reaches approximately 900 Mbps.

With 100 active users, offered demand reaches 1.5 Gbps—greater than the illustrative cell capacity.

This does not mean every real 5G cell provides 1 Gbps. Actual capacity depends on spectrum bandwidth, frequency band, MIMO configuration, TDD pattern, interference, radio conditions, transport capacity, device capabilities, and network configuration.

1784083522795

LinkedIn: :backhand_index_pointing_down: