It is quite interesting how the story of 5G is folding out. These days, conversations have come out of the speed factor and moved into different applications and use cases of 5G, ranging from specific devices to homes to entire cities. According to Gartner, worldwide 5G wireless network infrastructure revenue will reach $4.2 billion in 2020.
With such growing interest and adoption of 5G, it becomes imperative that service providers ensure a smooth experience for end users. Which is why testing keeping the end user in focus becomes a core and crucial part of the service. Being right in the midst of the testing world, we have experienced, and solved, the most critical test scenarios for 5G CPE. In a series of posts, we will reveal and discuss these scenarios.
In this first post, we will talk about an extremely popular component of 5G, the Throughput under different load conditions. For an end user connected over Wi-Fi to the 5G CPE, the throughput is decided by three factors:
- The Throughput of the 5G backhaul. For mmWave the external channel conditions will affect this throughput.
- The Wi-Fi client type (Wi-Fi 4, Wi-Fi 5, Wi-Fi 6) and the in-building(local) channel conditions. A suboptimal local Wi-Fi network will bring down the Throughput and this might get attributed to 5G. Clients connected over ethernet to the CPE does not get impacted by local channel conditions.
- The processing power of the CPE device itself (true for high client load, multi QoS traffic conditions).
As an example, if the CPE has a large number of clients trying to connect with an incorrect login credential or security configuration, the processing power available for the rest of the traffic reduces and affects throughput. The way out of this is to divide the throughput testing into different parts:
I) Test the 5G only Throughput under different load conditions
Increase the connected clients from 1 to maximum number supported over ethernet and run throughput tests. This test also applies to office use cases where a number of clients may connect via Ethernet. The same can be repeated under different background load conditions (HD video, FTP etc).
II) Test the Wi-Fi only Throughput (if feasible)
If the CPE architecture allows Wi-Fi only traffic testing it is highly recommended to do so. If the Wi-Fi is the weak link in the overall system this test helps to find and rectify issues. It also benefits to ensure that 5G doesn’t get the blame for any possible issues on the Wi-Fi part.
III) End-to-End Throughput
This is the most important throughput test because it is what the end customer will really experience. This test can be repeated for different client loads, background traffic conditions and time duration. Usually, this throughput will be lesser than the 5G throughput numbers in real environments.
UL UDP
|
DL UDP | UL TCP | DL TCP | |||||||||
# of clients | Wi-Fi only | 5G only | End-to-End | Wi-Fi only | 5G only | End-to-End | Wi-Fi only | 5G only | End-to-End | Wi-Fi only | 5G
only |
End-to-End |
1 | ||||||||||||
10 | ||||||||||||
30 | ||||||||||||
60 | ||||||||||||
120 | ||||||||||||
Max # of clients |
- 10 – 30 clients (common in homes)
- 60 – 120 clients (common in offices)
- 120+ clients (for public spaces)