In this multi-part series, I will continue comparing HPE Aruba Instant, Ruckus Unleashed and Ubiquiti UniFi WiFi 6 Access Points for home use. This time I am taking a look at performance difference including throughputs and range.
Setup
Since I only have couple WiFi 6E capable client device right now, and one of them is wired, I’ve decided to stick with WiFi 6 access points. This helps the cost for enterprise graded gears, especially on the second hand market like eBay or Amazon refurbished. For the detail cost comparison, please refer to the separate write up.
Specifically, I am comparing following models from each vender.
- UniFi – U6 Professional
- Ruckus – R750 and R350
- Aruba Instant – AP- 535
| Ruckus | Aruba | Ruckus | UniFi | |
|---|---|---|---|---|
| R750 | AP-535 | R350 | U6 Pro | |
| WiFi Version | WiFi 6 | WiFi 6 | WiFi 6 | WiFi 6 |
| 5 GHz | 4×4 | 4×4 | 2×2 | 4×4 |
| 5GHz 160 MHz | ⭕️ | ⭕️ | ❌ | ⭕️ |
| 2.4 GHz Wifi version | WiFi 6 | WiFi 6 | WiFi 6 | WiFi 6 |
| 2.4 GHz | 4×4 | 4×4 | 2×2 | 2×2 |
| Tx Power 2.4GHz (dBm) | 26 | 24 | 23 | 22 |
| Max Antenna Gain (dBi) | 3 | 3.5 | 3 | 4 |
| EIRP | 29 | 27.5 | 26 | 26 |
| Minimum Receive Sensitivity (dBm) | -102 | -96 | -101 | ? |
| Tx Power 5GHz (dBm) | 28 | 24 | 23 | 26 |
| Max Antenna Gain (dBi) | 3 | 5.4 | 3 | 6 |
| EIRP | 31 | 29.4 | 26 | 32 |
| Minimum Receive Sensitivity (dBm) | -102 | -93 | -101 | ? |
All four models are WiFi 6 capable. Other than R350, they all have 4×4 MIMO support on 5 GHz band.
Perfomance Comparison
In this category, I will compare followings:
- Max throughput
- Rise Time
- Ranged throughput
- Area of coverage/range
- Density Throughput Comparison (R750 and AP 535 only)
Unless otherwise specified, following setup is used.
- Client: Iphone 14 Pro Max (2×2 WiFi 6 device)
- AP transmission power: Auto or max.
Max throughput
Setup/Method
- Client device position: <5ft from the AP with full line of sight
- Test method: iPerf with basic command iperf -c [server ip] and iperf -R -c [server ip]
- Each iPerf tests were ran at least twice to reduce random variability. If one data is significantly different from the other, repeated the process until the data appear to stabilize.
- Picked better result of the two*.
Results
| (In Mbps) | U6 Pro | R750 | R350 | AP 535 |
| Client # | 1 | 1 | 13* | 1 |
| Min | 139 | 748 | 762 | 735 |
| Average | 622 | 760 | 790 | 763 |
| Max | 749 | 768 | 799 | 791 |
| WiFi Efficiency | 52% | 63% | 66% | 64% |
| (In Mbps) | U6 Pro | R750 | R350 | AP 535 |
| Client # | 1 | 1 | 13 | 1 |
| Min | 105 | 829 | 756 | 786 |
| Average | 659 | 861 | 874 | 813 |
| Max | 764 | 874 | 886 | 816 |
| WiFi Efficiency | 55% | 72% | 73% | 68% |
*R350 had 13 clients connected when the highest throughput number was observed, which on the different day than the original measurement, but since this was higher than the original data without any other client connected, I’ve decided to use this number instead.
Discussion
One of main metrics I use to assess the quality of WiFi Access point is the WiFi efficiency.
WiFi efficiency = Actual throughput/wireless link speed (PHY)
Since the PHY is essentially defined by WiFi standard, MIMO, and actual signal strength, it is the same across all access points at close proximity, but the practical throughput differs due to internal hardware and software optimization on the access points.
Normally, WiFi efficiency is said to be in the range between 50-70%. However, in the past consumer gears and UniFi are all fell more in the range of 50-60%. In fact, the latest Ubiquiti UniFi’s official site says the expected efficiency is 50-60%.
U6 Pro perfectly falls under the range. However, both Aruba Instant and Ruckus Unleashed are well above. Particularly, the download throughput on both Ruckus AP exceeded 70%.
Rise Time
Setup/Method
- Client device position: <5ft from the AP with full line of sight
- Test method: iPerf download command (iperf -R -c [server ip])
- Look at time it takes the throughput to reach its first peak.
Results
| U6 Pro | R750 | R350 | AP 535 | |
| Initial Throughput (Mbps) | 139 | 839 | 752 | 772 |
| First Peak Throughput (Mbps) | 703 | 862 | 868 | 797 |
| Time to peak (sec) | 5 | 2 | 1 | 1 |
Discussion
Both Ruckus and Aruba access points have essentially instant rise time. Even though R750 looked to have taken a second longer than R350 or AP 535 to reach its first peak throughput, it started out faster than any other AP so we could argue it actually took 0 to reach peak.
Rise time concept is indeed the biggest learning point for me when I first compared Ruckus Unleashed setup to UniFi. From what I can recall, UniFi AP had always have slow steady rise in its throughput over the first a few seconds as shown on current test.
So why do we care rise time?
Rise time is a common concept in electronics, which is analogous to acceleration time in cars. I think the shorter rise time translates into the faster data transfer during small file transfers.
In fact, I have experienced this myself. Whenever I transfer multiple 50-100 MB size files from local NAS server to WiFi client, there were always a few seconds of stall between each file transfer to start with UniFi setup. However, with Aruba and Ruckus setups, new file download starts immediately when the other one finishes.
At one point, I wondered if this had some effect on smoother, faster web page loading, but this could be more of subjective, perhaps placebo effect. If its real, then there maybe some other factor than Rise Time since most web page are fairly small in their file size.
Ranged throughput
Setup/Method
- AP position: R350 facing South. Other three were placed the other end of stairs one at a time, facing North at the same height, same orientation.
- Client device position: ~30ft from the AP with full line of sight
- Test method: iPerf with basic command iperf -c [server ip] and iperf -R -c [server ip]
Results
| U6 Pro | R750 | R350 | AP 535 | |
| Client # | 1 | 1 | 1 | >1 |
| Min | 156 | 391 | 491 | 546 |
| Average | 426 | 462 | 521 | 609 |
| Max | 499 | 520 | 551 | 647 |
| WiFi Efficiency | 36% | 39% | 43% | 51% |
| U6 Pro | R750 | R350 | AP 535 | |
| Client # | 1 | 1 | 1 | >1 |
| Min | 117 | 702 | 560 | 751 |
| Average | 613 | 731 | 642 | 789 |
| Max | 718 | 741 | 708 | 801 |
| WiFi Efficiency | 51% | 61% | 54% | 66% |
*I forgot to write down # of client devices connected to AP 535 when the ranged test was done, but as I recall there weren’t that many connected, and believe it was only couple more.
Discussion
It’s important to note that WiFi Efficiency calculated here used the max possible wireless link speed, which is 1200 Mbps rather than the real link speed at the 30 feet position. Using this as a reference made easier for relative comparison to max throughput.
Here I saw Aruba AP 535 outperformed all the others in both upload and download throughputs. U6 Pro was consistently the lowest on both upload and download.
One interesting result was that the AP 535 outperformed R750. This is because R750 had better signal strength than AP 535. R750’s throughput especially download was very impressive, but in this particular test AP 535 essentially maintained near peak throughput.
Area of Coverage/Range
Setup/Method
- Client device position: <5ft and 30ft from the AP with full line of sight
- RSSI measured using AirMac app
- Compatible ranged throughput comparisons in my home
- 2 AP setup ranged comparison between R750 x 2 and AP 535 x 2 for entire home
Results
| U6 Pro | R750 | R350 | AP 535 | |
| Tx Power | 26 | 28 | 23 | 24 |
| EIRP | 29 | 32 | 26 | 29.4 |
| 5ft | -30 | -25 | -29 | -25 |
| 30ft | -54 | -46 | -56 | -56 |
| dBm Loss | 24 | 21 | 27 | 31 |
| (in Mbps) | U6 Pro | R750 | AP535 |
| Min | 0 | 182 | 46 |
| Average | 146 | 235 | 158 |
| Max | 251 | 323 | 236 |
Discussion
RSSI is the signal strength seen by the client device. Not surprisingly, it highly depends on the transmission power from the access point, as well as other factors such as receptor sensitivity, antenna polarization etc.
It is important to note that dBm is a logarithmic scale unit, so difference in 3 dBm means twice the power level, 6 dBm is 4 times and so on.
From straight line of sight RSSI comparison, although it was expected, it’s quite impressive that R750 still retained very high RSSI at 30ft mark. Based on dBm difference, it has 10x more signal strength power to the client device at this spot than R350 or AP 535.
Storage in basement is where I put my network rack, NAS servers. But this area is surrounded by concrete walls. So it is one of the area in my home that’s WiFi unfriendly. So usually WiFi signal is best transmitted from the level above located AP.
With all 3 APs placed on the same spot, same orientation, Ruckus R750 maintained solid connection with very respectable throughput. AP535 also was able to show relatively high average throughput but the speed varied quite bit during the 30 seconds iPerf testing.
As for U6 Pro, its average number is misleading. It looks as if the ranged throughput in this area was matching the capability of AP535. However, as shown by the minimum throughput number of 0, it kept losing WiFi connection in this area. This persisted even after couple additional measurements. So practically, the WiFi connection at spot with UniFi was unreliable.
Density Throughput
Even though both Aruba Instant and Ruckus Unleashed APs showing phenomenal performance, I remember Ruckus sponsored and designed test comparing specifically R750 and AP 535 for high density setting, and concluding Ruckus is vastly superior in the setting. Although I can’t do anything close to the real density test, I thought I might throw a pseudo-density test to see if I can at least see their conclusion is not readily rejectable.
Setup
- Client device to AP distance ~10 ft
- Both AP were placed at the same spot
- iMac playing online stream with Airplay casting on Apple TV across the room
- All measurements were repeated at least twice to evaluate relative stability
- Max Swing is a dataset that had the biggest variation between min and max throughput
Results
| (In Mbps) | R750 | AP 535 |
| Client # | 28 | 30 |
| Min | 610 | 204 |
| Average | 644 | 515 |
| Max | 715 | 667 |
| WiFi Efficiency | 54% | 43% |
| (In Mbps) | R750 | AP 535 |
| Client # | 28 | 30 |
| Min | 638 | 512 |
| Average | 736 | 553 |
| Max | 779 | 571 |
| WiFi Efficiency | 61% | 46% |
| (In Mbps) | R750 | AP 535 |
| Client # | 28 | 30 |
| Up vs. Down | Up | Up |
| Min | 117 | 189 |
| Average | 432 | 441 |
| Max | 659 | 670 |
| WiFi Efficiency | 36% | 37% |
Discussion
First, it’s important to admit this is not even close to scientifically sound high density throughput test. The real high density throughput test requires multiple client devices not only connected but should be performing active data transfers simultaneously. They need to be high enough to stress the systems. If anyone is interested, the only data I can see is Ruckus’s own funded, designed comparison study published by Packet 6.
On my non-scientific test, I certainly didn’t put enough stress to the neither AP but the difference throughput between two appears relatively significant. This is not necessary validating Packet 6 test, but I could not invalidate it, and the result I am seeing can support with their test result.
Limitation
Despite test like throughput provides objective data, these data have a significant variation due to the multifactorial nature of WiFi performance. The things that can affect the performance range from WiFi interference, underlying network congestion, AP’s internal processing timing and others. Despite I tried to put a reasonable effort to match confounding factor variables, I did not stop or restrict my home WiFi use during these tests as that would not go well with my family.
In fact, I have certainly noticed variations of test results when I measured things different day, different timing.
For example, when the original max throughput on R350 were measured along with other APs, it did not do so well on the upload despite it was only serving a single client. Later on a different day, when I measured its throughput at the same location, I got higher throughput result despite it had 12 other clients being connected.
Another example is when changing wired network path from AP to NAS, the throughput changed counterintuitively.
With full end to end 1 Gbps connection, Ruckus APs fairly consistently outperformed Aruba AP-535 with Ruckus having multiple upper 800’s iPerf runs while Aruba’s best is lower 800’s. When the wired connection path is changed to include 10 Gbps path, no AP could reach 800’s. Then more importantly, Aruba AP-535 actually had more consistent, slightly higher throughput than Ruckus APs.
These wo examples confirms variability of the test results that are beyond AP itself, and they can even affect relative performance difference amongst different APs. So the message here is do not take my numbers as absolute.
Summary
In summary my key interpretation of these tests result are:
- Enterprise APs performance is objectively superior to U6 Pro including significant difference in rise time.
- Ruckus Unleashed R750 has higher range than others
- Though both Aruba and Ruckus APs WiFi efficiency are very impressive, when proper condition is met, Ruckus APs should have higher maximum WiFi Efficiency than AP 535.
- Ruckus R750 is suggestive of having better high density throughput than AP 535
But how do these results really translate into end user experience, especially as a home user?
U6 Pro vs. Others
I believe most home user will be able to appreciate the performance difference between U6 Pro and other 3 enterprise APs. However, the actual value of the benefits seen from these difference would vary amongst users.
For example, I have certainly noticed faster multiple small consecutive file transfer between the local NAS server to WiFi client device primarily due to the fact enterprise APs have no wait between files, while U6 Pro had at least 2-3 second before each file starting to download.
But if I transfer only 10 files, that’s merely 20-30 seconds difference a day as compared to 100 files, which would still be max 5 minutes difference. I will take a closer look into cost performance comparison in the future episode, but you can see the value really comes down to one’s usage and how much one think the time is worth.
Aruba vs. Ruckus
Despite Ruckus APs are objectively superior to the Aruba Instant AP 535 in most performance, how this translates into end user experience is much harder. First the difference is relatively small. Second, in some case AP 535 can outperform Ruckuas a home user we are unlikely to be pushing these systems to their limit in normal daily use.