Poor wifi throughput compared to Netgear R6300v2 router repurposed as AP
This thread has been locked for further replies. You can start a new thread to share your ideas or ask questions.
I'm looking to replace my Netgear R6300v2 repurposed from being my gateway router to an AP but I'm seeing some odd WiFi throughput that I'm trying to understand. Both it and the EAP225 are about 20 feet away from my laptop with no obstacles in the path and both are using channel 36 and 80MHz channel width. Windows 10 shows a link rate of 526.5 with either AP.
Measurements are done using iperf3 with default settings and with a server at the far end. Laptop has an Intel AC8265 which is a 2x2:2 card. Baseline throughput is about 920Mbps in either direction running over wired Gigabit Ethernet.
I administratively shut off the radio not being actively tested.
I measure 210-220Mbps in either direction using the Netgear and only half that with the EAP225.
What settings should I be looking at to debug this?
P.S.: InSSIDer shows -44dBm from the laptop with the EAP225 and -56dBm from the Netgear
- Copy Link
- Subscribe
- Bookmark
- Report Inappropriate Content
I notices a similar thing yesterday with my limited test runs with iPerf3 on EAP-245v3. One client hardwired to a gigabit switch (Cisco SG300) the other was a connecting wirelessly and was about 5-10m away from the AP.
My test results were as below:
TCP - Maxed out at around 140-160Mbps
UDP - Maxed out at around 45-55 Mbps. If i push 100Mbps, there's about 40-70% loss factor in most cases.
I'll be running some tests again this coming tuesday at a differnt sote installation where I would have some extra time run these throroughly. Will post an update here.
- Copy Link
- Report Inappropriate Content
What wireless client were you using when you got 140-160Mbps throughput?
Based on some things I read about another vendor's AP, I turned wireless encryption off but the throughput remains the same.
Tried using a laptop with an Intel AC7260; no difference
Tried using two laptops simultaneously as iperf3 clients; aggregate bandwidth is still around 100Mbps
- Copy Link
- Report Inappropriate Content
My Wireless client is an Intel 8265 NIC laptop (i7-7820hk/16GB RAM and SSD). What I am not sure is wired client - it had an AV which I didn't get a chance to deactivate.
I'll be using a linux based bootable drive on tuesday and hopefully be able to test this much more thoroughly..
- Copy Link
- Report Inappropriate Content
I have the same NIC but you're getting higher TCP throughput. You wouldn't think EAP225 vs EAP245 would make a difference since the client is 2-stream but maybe beamforming does have an effect!
Edit: I went through each network element carefully and discovered that the 2m long Ethernet cable from the PoE injector to Gigabit Ethernet switch was allowing only a Fast Ethernet connection! Now I'm at 150-160Mbps second throughput too but need to make further progress.
- Copy Link
- Report Inappropriate Content
my PoE injector included with EAP225 V3 negotiates 1,000 Mbps FD with my switch. Check your cable again. Is it shielded?
As for iperf performance, see my post #5 in this thread for a detailed measurement of EAP225 throughput/goodput.
Long story short: The EAP225 achieves full wireless goodput of ~610 Mbps up to 670 Mbps in ideal situations (this means at maximum WiFi rate of 867 Mbps in the 5 GHz band).
In 802.11ac mode @80 MHz channel width max. wireless bandwidth is 867 Mbps. Max. wireless throughput is ~700 Mbps (that's minus ~19% for WiFi protocol overhead) and TCP/IP goodput is ~640 Mbps on average (that's ~8.5% of the WiFi bandwidth = ~7% of total bandwidth for TCP protocol overhead or in other words: ~26% protocol overhead for TCP and WiFi together).
- Copy Link
- Report Inappropriate Content
Thanks for the info. What baffles me is that you were able to push 720Mbps of UDP traffic with minimal to 0 loss whereas if I wasn't able to push beyond 50Mbps without a loss factor. I'm guessing this is due to the antivirus that was on the othe PC connected via LAN (wasn't able to disable it).
The other issue is that the WiFi space is extremely crowded, 6 storey building in the middle of a major city. When I do a rogue AP scan I get 100 entries :(
- Copy Link
- Report Inappropriate Content
I should have been clearer above, but I did replace the cable and got a gigabit connection at least per the lights on the switch.
With your numbers in hand, I redid the experiment by moving closer to the AP so that the PC reported a 867 PHY rate. There is no one else on that channel.
iperf3 got close to 300Mbps over TCP which is half what you report.
The laptop acting as the iperf3 server is wired to the same switch as the EAP and I can get full throughput if the laptop being used as a client uses a wired connection to that switch.
Here's the output of ifconfig:
$ ifconfig -a
ath0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:1664591 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath1 Link encap:Ethernet HWaddr D2:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:4564 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath10 Link encap:Ethernet HWaddr D6:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2301736 errors:0 dropped:0 overruns:0 frame:0
TX packets:856136 errors:228 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3298060014 (3.0 GiB) TX bytes:649636858 (619.5 MiB)
ath11 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath12 Link encap:Ethernet HWaddr D2:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath2 Link encap:Ethernet HWaddr D6:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:5666 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
br0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
inet addr:192.168.1.4 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1697016 errors:0 dropped:164757 overruns:0 frame:0
TX packets:34714 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:194263230 (185.2 MiB) TX bytes:12188604 (11.6 MiB)
eth0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2398349 errors:0 dropped:10 overruns:1 frame:0
TX packets:2335268 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:829119364 (790.7 MiB) TX bytes:3310122358 (3.0 GiB)
Interrupt:4
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
wifi0 Link encap:UNSPEC HWaddr CC-32-E5-74-B5-E4-00-00-00-00-00-00-00-00-00-00
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:1602014 errors:0 dropped:0 overruns:0 frame:0
TX packets:2539082 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:309495070 (295.1 MiB) TX bytes:529322634 (504.8 MiB)
Interrupt:47 Memory:b8100000-b8120000
wifi1 Link encap:UNSPEC HWaddr CC-32-E5-74-B5-E5-00-00-00-00-00-00-00-00-00-00
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:2405146 errors:26540 dropped:0 overruns:0 frame:0
TX packets:2896092 errors:228 dropped:352 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:3384172840 (3.1 GiB) TX bytes:673333712 (642.1 MiB)
Interrupt:40 Memory:b2000000-b2200000
- Copy Link
- Report Inappropriate Content
Hi @unmesh,
There is a bug in the current EAP225V3 firmware (2.7.0) where connection rates inexplicably fall from 866Mbit/s to ~400Mbit/sec after some period of time after an AP reboot. This appears to be a new bug that was introduced in 2.7.0. It was reported back in early March and TP-Link. They claim it will be fixed in the next release.
Try restarting the AP to see if this helps. Also try changing the channel and order toggling between 20/40/80 and 80 MHz BW.
I have 2xEAP225V3, 1xOC200, and Fios Gigabit. Be sure to try different speed test websites/apps on different devices. I find that many speed tests are very reliable above a few hundred megabit. Ookla, Fast.com, Speedtest.net, dslreports, etc. often given varying results and can give different results at different times of day.
You should be seeing ~300-500 Mbit/sec DL from an Ethernet connected EAP225V3. I usually see a bit higher on the UL (~400-550Mbit) for some reason. But during the day, when the internet is busy, the DL speeds can drop to ~200-300 Mbit/sec. I have not tested with iPerf, but I do use inSSIDER quite a bit. Also, for Win10, Windows Apps (e.g. Ookla from the Microsoft App store) seems to perform much better more reliably than running it in a browser (e.g. Chrome). But in general, I find that running apps from fast modern mobile devices (like my iPhone XSM) tend to give more consistent speed test results.
Set your AP's for different channels, and lower the transmit power to medium. Disable band steering, use separate SSID's for 2.4 and 5.8, and disable Airtime Fairness.
Some on this board may have varying options on "Goodput" (that is Application level throughput) as a proportion of Wifi link rates. But in general, I find that useable bandwidth across Wifi 3rd (g), 4th (n), and 5th (ac) generations is pretty typically 40-60% of the link rate. So, if your link speed is 600 Mbit/sec you should be seeing ~300 Mbit/s on various speedtests.
Also double check / swap your ethernet cable (try the one going to the Netgear) to see if it help.
-Jonathan
- Copy Link
- Report Inappropriate Content
I plan to do some more experiments today and will include your suggestions especially since I have lots of Ethernet cables!
Also, the bug report is intriguing. Does TP-Link allow you to go back to older firmware versions?
Since I was worried about the number of drops reported in the ifconfig output, I did a 900MB file upload after a reboot and there are very few errors/drops though the transfer rate to my local server averaged about 20MB/s.
ath0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:4927 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath1 Link encap:Ethernet HWaddr D2:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:10 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath10 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath11 Link encap:Ethernet HWaddr D2:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
ath12 Link encap:Ethernet HWaddr D6:32:E5:74:B5:E5
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:651535 errors:0 dropped:0 overruns:0 frame:0
TX packets:94873 errors:35 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:977764411 (932.4 MiB) TX bytes:9892672 (9.4 MiB)
ath2 Link encap:Ethernet HWaddr D6:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:8 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
br0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
inet addr:192.168.1.4 Bcast:192.168.1.255 Mask:255.255.255.0
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:5529 errors:0 dropped:894 overruns:0 frame:0
TX packets:150 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:673173 (657.3 KiB) TX bytes:42847 (41.8 KiB)
eth0 Link encap:Ethernet HWaddr CC:32:E5:74:B5:E4
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:92442 errors:0 dropped:0 overruns:0 frame:0
TX packets:651608 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:7776534 (7.4 MiB) TX bytes:977797274 (932.5 MiB)
Interrupt:4
lo Link encap:Local Loopback
inet addr:127.0.0.1 Mask:255.0.0.0
UP LOOPBACK RUNNING MTU:16436 Metric:1
RX packets:0 errors:0 dropped:0 overruns:0 frame:0
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:0
RX bytes:0 (0.0 B) TX bytes:0 (0.0 B)
wifi0 Link encap:UNSPEC HWaddr CC-32-E5-74-B5-E4-00-00-00-00-00-00-00-00-00-00
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:3816 errors:0 dropped:0 overruns:0 frame:0
TX packets:5677 errors:0 dropped:0 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:689989 (673.8 KiB) TX bytes:1179052 (1.1 MiB)
Interrupt:47 Memory:b8100000-b8120000
wifi1 Link encap:UNSPEC HWaddr CC-32-E5-74-B5-E5-00-00-00-00-00-00-00-00-00-00
UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1
RX packets:664312 errors:5060 dropped:0 overruns:0 frame:0
TX packets:99532 errors:35 dropped:4 overruns:0 carrier:0
collisions:0 txqueuelen:1000
RX bytes:994991098 (948.8 MiB) TX bytes:9937304 (9.4 MiB)
Interrupt:40 Memory:b2000000-b2200000
- Copy Link
- Report Inappropriate Content
unmesh wrote
iperf3 got close to 300Mbps over TCP which is half what you report.
Maybe, your wireless NIC is SISO or you have additional SSID on this AP with clients acquiring too much AirTime, so not much is left for the stream in the channel you use for testing. Or interferences from nearby offices slow down things.
My office is amongst residental buildings, no other 5 GHz APs visible indoors (that's the advantage of having indoor-only channels). I measure between a MacBookPro (UNIX) and a Dell server (Linux). Distance to the EAP is ~2m, free line of sight, wall mounted 1m above the floor.
I use iperf or iperf3 b/c I want to measure the pure WiFi throughput, but not Internet, not router, not web server running »speed tests« throughput. iperf is the most precise tool I'm used to since it was created.
If I want to measure Internet throughput, I use iperf on our public Internet servers in various data centers.
Only for Web browsing l would use web-based »speed« tests (they actually measure throughput, not speed).
Ookla, dslrepots etc. vary greatly in results, b/c of high latencies of foreign networks (all dslreports EU servers are currently down), or b/c of overloaded servers or temporarily overloaded peering points.
- Copy Link
- Report Inappropriate Content
Information
Helpful: 0
Views: 5582
Replies: 28
Voters 0
No one has voted for it yet.