|
My kids' school (Yes, I am the computer guy of the school as a part-time job hobbyist  ) has a mobile cart with 16 strictly identical hp laptops, that is used by the teachers in their classrooms. The lappies are connected to two wireless access points (1 access point for 8 laptops) placed inside of the cart, which is itself connected to the school network via cable. The school has a server and its own domain.
Problem: At most, 6 to 8 laptops can connect at the same time without interruption or loss of the wireless connection, for surfing the internet or opening word documents from the server). When more laptops are used, navigation on the network (files or internet) gets really slow or even freezes.
I lack experience in mass wireless technology, so maybe you guys can help; What I tried so far :
- putting all 16 laptops to the network via a switch and cables instead of wireless, I get an average of 50Mbit/s on each, everything runs fine, so I suppose everything is ok from the server point of view.
- access points are 300Mbit/s capable, which means a theoretical 150Mbit/s (wireless is not synchronously bidirectional) for 8 laptops, or about 20MBit/s for each of them, in an ideal world. But I get between 20Kbit/s and 4Mbit/s at the very best. The wifi network card of the laptop is 2.4GHz.
- I separated the channels (used chan 1 and chan 6 respectively on the access points), no improvement
- I used 5GHz access points with 5GHz wifi usb dongles instead of the 2.4GHz ones, with almost no improvement.
Am I hitting a physical limitation of the Wifi bandwidth when trying to use 8 machines on a single access point ? Is there any different type of access points that are designed to handle more connection or with increased bandwidth ?
Any idea of what else could go wrong ?
|
|
|
|
|
- Are the access points in a metal cabinet? Is the cabinet closed while the computers are in use?
It could be that you are not actually connecting to the APs in the cabinet, but to different APs much further away. - [EDIT] Check the channel width on the access points (using the web interface). It should be 20MHz, not 40MHz.
A channel is 5MHz wide, but 20MHz are required for Wi-Fi. This means that a 20MHz channel covers a width of 5 channels, i.e. "channel 6" at 20MHz actually uses channels 4 to 8, and "channel 1" at 20 MHz actually uses channels -1 to 3.
If you try using a 40MHz channel, you have "channel 6" actually using channels 2 to 10, and "channel 1" using channels -3 to 5, overlapping the channels used by "channel 6". - Connect one laptop to each access point (on different channels - 1 & 6), test the throughput - upload only, download only. You should see a high throughput on both laptops.
- Get a Wi-Fi network sniffer (there are freeware programs that you can install on your phone), and see if there are other networks using the same channels. For example, the school may have a Wi-Fi network for the teachers, for administrative stuff, etc.
The issue may be the number of transmission collisions. When there is a collision (two devices attempting to transmit at the same time), the devices retry after waiting a while. The more devices you have on the network, the more collisions are likely. If each device is attempting to transmit at the maximum rate (150 Mb/sec), you are attempting to send 1350 Mb/sec (don't forget the Access Point). Collisions are guaranteed under this case, with the associated retries.
Freedom is the freedom to say that two plus two make four. If that is granted, all else follows.
-- 6079 Smith W.
modified 17-Feb-21 17:12pm.
|
|
|
|
|
Daniel Pfeffer wrote: Get a Wi-Fi network sniffer
Good idea - I use https://play.google.com/store/apps/details?id=com.vrem.wifianalyzer[^] which is brilliant for a free app - open source as well.
"I have no idea what I did, but I'm taking full credit for it." - ThisOldTony
"Common sense is so rare these days, it should be classified as a super power" - Random T-shirt
AntiTwitter: @DalekDave is now a follower!
|
|
|
|
|
|
Thanks for the suggestions and the time you took to answer  .Daniel Pfeffer wrote: Is the cabinet closed while the computers are in use?
Only wood. I took the APs out and put them in separate corners of the room without any improvement.
Daniel Pfeffer wrote: Check the channel width on the access points
I tested all possible configurations : overlapping chans, no overlapping chans, 20MHz, 40Mhz. Best results with chan 1 and chan 6 or 11, 20MHz, but still not sufficient.
Daniel Pfeffer wrote: You should see a high throughput on both laptops
With 2 laptops, 2 access points, I have up to 6Mo/s (tried 2 different access points of different brands that are respectively 300Mb/s and 500Mb/s capable), all good. It starts to go down when I have more than 3 laptops on one AP. Two bizarre things here : why only 6Mo/s where it should be at least 20, and why the breakdown when 4+ are connected...
Daniel Pfeffer wrote: Get a Wi-Fi network sniffer
Done as well, and most other chans are indeed used by the school neighbourhood, but with very, very little signal strength, so they are barely visible.
Daniel Pfeffer wrote: The issue may be the number of transmission collisions
Interesting. How can I avoid this ? Could load balance of limitations per device help ?
|
|
|
|
|
Rage wrote: Daniel Pfeffer wrote: The issue may be the number of transmission collisions
Interesting. How can I avoid this ? Could load balance of limitations per device help ?
If you can't switch to 5ghz (lots more channels available); about all you can do is see if you can reduce transmit power on the access points and clients.
Wifi devices can only transmit 1 at a time on a channel [*]; meaning that counter-intuitively increasing transmit power in a crowded environment only makes things worse (you're seeing more devices transmit, so each devices share of the transmit time goes down).
[*] not strictly true in the latest versions of the wifi standard if every device in range is using it; in maybe 10 years that's the case this might actually be relevant. (But probably not because someone will have a cheap #InternetOfShit device on the network that saved $0.01 by using a 20 year old wifi module.)
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, weighing all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|
Thanks.
The more I look at it, the more I think this setup with 16 pcs on wireless cannot work properly.
|
|
|
|
|
You need better than cheapo wifi-N 2.4ghz anyway.
I hope the laptops are capable of doing better and it's just the access point that's an obsolete paperweight.
Did you ever see history portrayed as an old man with a wise brow and pulseless heart, weighing all things in the balance of reason?
Is not rather the genius of history like an eternal, imploring maiden, full of fire, with a burning heart and flaming soul, humanly warm and humanly beautiful?
--Zachris Topelius
Training a telescope on one’s own belly button will only reveal lint. You like that? You go right on staring at it. I prefer looking at galaxies.
-- Sarah Hoyt
|
|
|
|
|
There are now 4 access points, 2 of them are 5GHz capable, and new. I am not sure about what features make an AP better than another one.
The hp laptops are only 2.4GHz capable, but I tried with 5GHz wifi usb dongles, to no improvement (in a setup with 1 5GHz AP and 4 laptops, the fourth laptop browser froze while they were trying to access the internet at the same time).
|
|
|
|
|
|
That is what I had in mind as next step, but I was not sure if physically the Wifi at 2.4Ghz really can handle more than 8 clients - It is worth a try though, thanks.
|
|
|
|
|
It's probably a built in limitation in the access point.
Both at home and at work I have Unifi access points, they have a limit of 127 clients per radio (so 254 in total), that should be enough. 
|
|
|
|
|
The current APs are also 254 clients capable, but this apparently does not mean that they guarantee bandwidth.
A 300Mbit/s device = 37Mo/s one direction = 18Mo/s both directions, so 254 devices would get with 100% efficiency about 70ko/s each, which is ... quite slow. 
But 18Mo/s for 8 clients is still about 2Mo/s/client, which should be sufficient for accessing a file or even watching SD videos from the internet, so I do not understand why I run into problems.
|
|
|
|
|
Bandwidth balancing certainly isn't built in to all access points.
This is a problem when using the hub model, clients are greedy, so when they receive a package they immediately send the next request allowing no room for other clients to use the bandwidth.
So if your AP doesn't have bandwidth balancing, it's the first to come that wins, the rest gets a timeout after some retries.
There are three solutions that I can see.
A) Get an AP with bandwidth balancing
II) Change the settings for TCP retries and/or timeout.
3) Block YouTube in the DNS.
|
|
|
|
|
Thanks, Jörgen, I'l ltry this  . Two of the four available APs can be configured.Jörgen Andersson wrote: lock YouTube in the DNS.
|
|
|
|
|
|
Here is another open-source monitoring tool: Cacti[^]
|
|
|
|
|
What do you see when you connect to the routers' admin page? Just those laptop connections, or everyone else in the place with a phone connecting to the free WiFi?
|
|
|
|
|
Only the laptops (and there is no free Wifi possibility - and even if, the Wifi is only activated when the cart is in use, so that would be a sporadic problem).
|
|
|
|
|
I've just found my first processor bug. No, not on one of the big guys like ARM or Intel but still a decent find on the Rabbit family of processors. The bug seems to have been there from the earliest days and to find it 20 years later is quite unexpected.
I've always said that I like programs you debug with an oscilloscope and hardware bugs you find with a test program, but this is the first bug on silicon and not at board level.
The feeling is quite ecstatic (what a big nerd am I!)
Mircea
|
|
|
|
|
But isn't this like "It must be a compiler bug!"
|
|
|
|
|
Almost, but I have a hierarchy of esteem where at the bottom are software hacks (like me), compiler writers are some levels above and chip designers are almost at the top.
To find a bug so high up the food chain is exhilarating
Mircea
|
|
|
|
|
Most probably not. Test again and again ask yourself again and again is it really the processor and not your side. Most probably you will find it is your side, this from experience from my side 
[Edit]
Otherwhise you can name the bug
modified 27-Mar-21 21:01pm.
|
|
|
|
|
Quote: Most probably not. Test again and again ask yourself again and again is it really the processor and not your side.
I did and I'm fairly sure this is not the case. Filed a bug report and we'll see what happens. I don't have any qualms admitting my errors so I promise I'll post the "denouement"
Mircea
|
|
|
|
|
Good luck then
modified 27-Mar-21 21:01pm.
|
|
|
|
General 
News 
Suggestion 
Question 
Bug 
Answer 
Joke 
Praise 
Rant 
Admin 
Submit an article or tip
