WiFi is slower with pfsense vs Untangle. Any thoughts?

[bokolobs]

Hi! I borked something after the update to 23.01 and took the chance to test Untangle 16.6.2. With the same vanilla configuration, I was getting 620 Mbps from my Omada AP with Untangle, vs 500 Mbps from the same AP with pfsense. I have a 1 Gbps up/down connection. I couldn't believe it at first, so I installed pfsense CE 2.6 and using stock configuration, tested my WiFi again. Went back to getting 500 Mbps from the same AP. I've always thought that getting 500 over WiFi was the best my AP could do. Any ideas on why this is the case? Is there a setting in pfsense I can change to achieve the 620 Mbps I'm getting in Untangle? Thanks in advance.

[Gertjan]

@bokolobs

The missing info is ... a lot.

What is the LAN port negotiated speed ?
What happens when you swap (assign) LAN and WAN ?
What is the driver name of the '500' interface ? The brand and type network of the network card involved ?
If applicable : https://forum.netgate.com/topic/177885/upgrade-to-23-01-wan-speed-halved/

From a pfSense point of view : your "Omada AP" is just another LAN device.

What details about the pfSense hardware ?

[bokolobs]

Hi @gertjan !

What is the LAN port negotiated speed ?

Negotiated speed of all interfaces is 1 Gbps.

What happens when you swap (assign) LAN and WAN ?

I haven't tried swapping them.

What is the driver name of the '500' interface ? The brand and type network of the network card involved ?

I'm using 4-port Intel I225-V(3) in a Qotom N5105 device with 128 GB M.2 NVME and 8 GB RAM.

From a pfSense point of view : your "Omada AP" is just another LAN device.

Yeah, that's what I was thinking as well. Really surprised at the difference in WiFi speeds. Also, I configured fq_codel on both with the same download and upload limits (800 Mbps). I'm getting consistent speeds using wired connection with Untangle. However, I'm getting about 15% less upload speed with pfsense.

[stephenw10]

How are you testing?

Can you get the full speed when using a wired connection?

Steve

[bokolobs]

@stephenw10

Yes! I get full speeds when wired. I tested using iperf and speedtest.net. However, when using fq_codel, I'm getting lower upload speeds. I did a cursory search online and found some people with the same experience, like this one. Now wondering if I should just turn off SQM.
Thanks.

[Patch]

@bokolobs said in WiFi is slower with pfsense vs Untangle. Any thoughts?:

I tested using iperf

Through pfsense or to/from pfsense? Running iperf on pfsense is known to cause spurious results.

[bokolobs]

@patch
I used the iperf package. From pfsense I get 940, to pfsense I get 920.

[stephenw10]

So between a local client and using pfSense as the iperf server?

It's always better to test through pfSense rather than directly to/from it if you can. pfSense is not a good server.

[bokolobs]

@stephenw10
Yes. I tested it again using an iMac directly connected to the pfsense box. The line auto-negotiated to 2500baseT. iperf using the pfsense box as server, I got 2.35 Gbps. Using -R, I got ~2.20 Gbps. So I guess in my case it's a good server?

In one area of our house where WiFi signal is weak, I could get ~80 Mbps using Untangle. Now, it won't even connect with pfsense. It's probably a problem with the combination of my appliance + FreeBSD + TPLink Omada. I also get the same speeds with OPNsense. I feel I'm leaving something on the table performance-wise with FreeBSD. I don't know what to do except I don't want to pay $150 for the Untangle license.

[stephenw10]

Hmm, that starts to feel more like the hardware you're running pfSense on is somehow affecting the wifi signal strength/connection errors. There's nothing the OS can do that would prevent a client even connecting to the AP. Is it physically close to the AP?

[bokolobs]

@stephenw10
Yeah, I’m about 2 feet away from the AP. Just to be absolutely sure, last night, I ran some tests using pfsense and then installed Untangle/Arista and ran the same tests. To me, at least for my particular hardware (box, switch, AP), the results are definitive.

Red - pfsense without traffic shaping/limiters
Yellow - pfsense with traffic shaping/limiters (650 up/down fq_codel limits)
Green - Untangle with traffic shaping/limiters (650 up/down fq_codel limits)
Blue - Untangle without traffic shaping/limiters

You might be right, something in my hardware + pfsense somehow makes my wifi connection weaker. This is a head scratcher for sure. Could it be because Omada also runs on Linux?

[stephenw10]

Hmm, hard to see how. Do you see the same latency in each setup?

[bokolobs]

@stephenw10 For the yellow results, jitter for both download and upload is less than 10 ms. But when speeds hit more than 600 in any of the results, jitter increases to ~15 ms, both for pfSense and Untangle. I really want pfSense to work like the green and blue results...

[rcoleman-netgate]

@bokolobs said in WiFi is slower with pfsense vs Untangle. Any thoughts?:

Yeah, I’m about 2 feet away from the AP

FWIW there is such a thing as too close to the AP.

[stephenw10]

Yup, that's true.

What I really meant there was how far is the pfSense hardware from the AP. Or the test client.
But since Untangle is being run on the same hardware (?) that shouldn't make any difference.

Steve

[SteveITS]

This is starting to feel a bit like the old story about the janitor who removed the network (cable) terminator during lunch to replace a missing pawn in their chess game. [just that it’s something weird and unexpected]

@bokolobs is there a switch between the AP and your router? If not I’d connect one and test from the iMac to wireless through the switch, to isolate from the router.

[bokolobs]

@stephenw10
Hi! AP is in another room separated from the router appliance by two concrete walls.

@steveits
Hi! Yes, there is an Omada PoE switch between the router and the AP. I tested the wired connection to the iMac two ways: directly connected to the router (router as server: ~2.35Gbps -R: ~2.20Gpbs); connected to the Omada switch (940/920 Mbps).

@rcoleman-netgate
Yeah. But I tested from the distance for both.

[bokolobs]

@stephenw10

Could it be related to the bios setup?

[SteveITS]

@bokolobs Does your AP have a (second) Ethernet port on it? Just thinking about testing wired through the AP. Because pfSense on a router can't affect a device-to-AP connection, where it's not in the middle. So it seems like you should somehow narrow down where the slowdown is happening.

All that said, wireless connections do change over time. My Windows PC doesn't move, yet the network properties window often shows a different link speed whenever I look at its wireless settings. I would think that sort of thing would be inconsistent though.

[bokolobs]

@steveits

Thanks! Unfortunately, my AP only has one port. I know WiFi can be inconsistent. What's boggling is I've done the tests 3 times now, using fresh install of 23.01 and Untangle 16.6.2, but the results are consistent. Which, to summarize, are two points: 1)I'm getting slower WiFi downloads and uploads when using pfSense vs Untangle, with or without traffic shaping and 2) I'm getting slower uploads when using wired connection AND traffic shaping, fq_codel limiters (650). When using wired connection without traffic shaping, I get the same download and upload speeds. It's driving me nuts.