Jump to content

Recommended Posts

Posted

Okay so I work at a brand new school (new build and refurb of the old). As part of the project the whole IT infrastructure has been replaced with something a little 21st century. As part of this a whole host of HP switches have been installed (which I do love) and a HP wireless system (which I'm not a fan of yet). All installed by the communication and data contractor, I had no say in what was purchased or installed.

 

Basically we have a HP MSM760 controller with 17 HP MSM460 AP's. The issue comes with the throughput of the WAP's, or rather than the combination of which device is connected to it. On my own HP ProBook 430 G1 the internet speed is solid as a rock on speediest.net, 20 up/20 down with a decent ping and a perfectly flat good connection bar on the speed test. Using the schools laptops (HP 350 G2), I can never seem to get a connection above 7 up/7 down or so with a terrible ping and a jagged connection bar.

 

Obviously if my personal laptop works perfectly, it can't be the WAP, right? Except even moving files over the wireless is only around 10MB/s. And the school has around 80 of the HP 350's, so I don't want to just say "it's the client device", even though they're not the best machines in the world. Heck, they even work perfectly on my 50mb Virgin Media connection at home!

 

Of course, all of the wired network infrastructure in the school works brilliantly as it should.

 

Does anyone have any ideas on what could be the cause of this?

 

Thanks in advance!

Posted
Sounds like classic firmware update time, if available (on the controller). So often issues like this are caused by bugs. Then look at the drivers on the laptops. Start simple.
Posted

I'd check/compare the connecting speed over wifi (probook/350), it maybe the 350 have less radios or lesser wifi support like 2.4 and 5ghz.

 

How did you test and was it completely controlled (if there are any other wifi devices in range you could get interference / degradation), wifi only supports 1 client at a time, and the air is the collision domain.

 

We recently brought a Wi-Spy/channelizer because we had no tool to help us migrate from coverage to density, and see how much saturation (air time) is happening.

 

You might also want to look at breaking your test down to just network performance by using something like iperf (https://iperf.fr/iperf-download.php). That should better reflect what your wifi is doing seperate to your inet speed.

Posted

i'd agree with 3s-gtech start simple, check for newer firmware / drivers

 

also check if there are any IPv6 issues (we had 30 new HP elitedesk 800 g1, which i217 intel cards spamming our network with ipv6 traffic)

Posted

Thanks for your help guys thus far.

 

Unfortunately, I updated the firmware on the WLAN controller and WAP's as soon as they got installed. In hindsight I wished I had tried it on the older firmware. Doh!

 

The 350 already has its latest drivers (as you tell I'm a bit fussy about keeping everything up to date). My moto is not 'If it ain't broke, don't fix it', but rather 'If there's a newer version, it must be better'.

 

So I took the 350 home with me last night to try there. At home I have a 50mb Virgin Media connection with a Ubiquiti UniFi Pro showering my house in Wi-Fi. It's almost the same - my iPhone, ProBook, iPad all pull down the max 50mb, but the 350 can only get down 30mb. While this is a problem that it can't do that, that 30mb is still more than what the school gets, but at school at 350 can only pull down around 5mb or so of 20mb!

 

The plot thickens...

 

- - - Updated - - -

 

Sounds like classic firmware update time, if available (on the controller). So often issues like this are caused by bugs. Then look at the drivers on the laptops. Start simple.

 

Thanks, I already covered this

Posted
I'd check/compare the connecting speed over wifi (probook/350), it maybe the 350 have less radios or lesser wifi support like 2.4 and 5ghz.

 

How did you test and was it completely controlled (if there are any other wifi devices in range you could get interference / degradation), wifi only supports 1 client at a time, and the air is the collision domain.

 

We recently brought a Wi-Spy/channelizer because we had no tool to help us migrate from coverage to density, and see how much saturation (air time) is happening.

 

You might also want to look at breaking your test down to just network performance by using something like iperf (https://iperf.fr/iperf-download.php). That should better reflect what your wifi is doing seperate to your inet speed.

 

I didn't delve into this too deeply, but both devices were connecting at 150mbps. The 350's don't support 5ghz, so I thought we may as well turn it off on the WAP's/controller, but it didn't seem to make any difference.

 

Again, it wasn't a 'controlled' test - we simply picked a classroom and set up some devices. I'll be the first to admit I'm not the most technical person when it comes to enterprise wi-fi (I prefer things like prosumer Ubiquiti UniFi etc.).

 

A colleague of mine has an Android tablet with this app the scans Wi-Fi channels and such like, might be a starting point, although we have just tried using the single AP with no other AP's in range from nearly homes/other WAP's, so unfortunately it doesn't necessarily sound like an interference issue.

 

After today the internet is being cut off for the best part of 5 or 6 weeks to move the connection to a temporary location where the school will be based whilst the build is being finished, so will absolutely need to look at something like iperf. Will play with it today, thanks!

 

Help much appreciated so far!

Posted (edited)

Okay, so here are my finding now I'm back on site and have brought my UniFi Pro access point from home. Hope you can see the images!

 

Just realised you can't see the speed:- with HP Wi-Fi the download speed was around 7Mbps, Lan 20Mbps, and UniFi around 9Mbps. Weird!

350 with HP Wireless.png

350 with LAN.png

350 with UniFi Wireless.png

Edited by exonetsystems
Posted
Attached is the reults from iperf. The top set of results are using the 350 directly plugged into the LAN, and the second set below is the results using the MSM460 AP with MSM760 controller. It's almost as if it's being throttled by the AP somehow, or maybe the wireless on this laptop really is that bad!

iperf.png

Posted
Having obviously spent a considerable sum on the infrastructure hardware, has the school considered engaging with a professional consultancy to unpick the network performance? A day or two would probably solve all your issues and you would learn a thing or two along the way.
Posted
Having obviously spent a considerable sum on the infrastructure hardware, has the school considered engaging with a professional consultancy to unpick the network performance? A day or two would probably solve all your issues and you would learn a thing or two along the way.

 

You would think so, but due to the politics of the build, the school aren't actually the client here - the government is. This project is already millions of pounds over and money is saved wherever possible. They had a chap come in to specifically configure the Wi-Fi on Monday, but he was here no more than hour and the only test was to bring up a webpage. A colleague of mine has gotten in touch but was met with a rude response of "it works" and won't come back to the site. Well yes, but not correctly. Psh! :mad:

Posted

To provide an update, I swapped the wireless card on my ProBook with the card on the 350. The 350 with the ProBook card in does seem to make the internet better (although still can't push/pull a solid 20mb connection). Obviously the ProBook with its original card in is solid as mentioned previously.

 

Just trying absolutely anything at the mo!

Posted

I was trying to think of the other information that was important yesterday and couldn't remember the name, today i found it MCS Index there's a good table on Demystifying Modulation and Coding Scheme (MCS) Index Values

 

This is the available speeds that wifi can connect at (spatial stream, channel band widths and Guard Intervals (wait time)). The other thing is power (which may be affected if laptop is plugged into ac or not), also if the MCS Index changes alot this can have negative impact on performance. I'd also check if you can see what the radio power levels are set. device manager -> wifi adapter -> properties -> advanced tab -> look for transmit power or power output. There's a bunch more settings there that maybe be able to be tweeked to give better performance.

 

I was interested to see here so I did a quick test between my probook (6550b) and a 215 which is a baby 350 (no 350s in my office) and both connected at 300mbs, top is probook (Broadcom 43224AG), bottom is 215 (bcm943228hmb)

 

Accepted connection from 10.105.89.10, port 2141

[ 5] local 10.105.89.112 port 5201 connected to 10.105.89.10 port 2142

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-1.00 sec 17.7 MBytes 148 Mbits/sec

[ 5] 1.00-2.00 sec 17.6 MBytes 147 Mbits/sec

[ 5] 2.00-3.00 sec 17.9 MBytes 150 Mbits/sec

[ 5] 3.00-4.00 sec 19.3 MBytes 162 Mbits/sec

[ 5] 4.00-5.00 sec 17.2 MBytes 144 Mbits/sec

[ 5] 5.00-6.00 sec 18.3 MBytes 154 Mbits/sec

[ 5] 6.00-7.00 sec 17.8 MBytes 149 Mbits/sec

[ 5] 7.00-8.00 sec 18.3 MBytes 153 Mbits/sec

[ 5] 8.00-9.00 sec 17.1 MBytes 143 Mbits/sec

[ 5] 9.00-10.00 sec 18.1 MBytes 151 Mbits/sec

[ 5] 10.00-10.03 sec 489 KBytes 174 Mbits/sec

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-10.03 sec 0.00 Bytes 0.00 bits/sec sender

[ 5] 0.00-10.03 sec 180 MBytes 150 Mbits/sec receiver

-----------------------------------------------------------

Server listening on 5201

-----------------------------------------------------------

Accepted connection from 10.105.88.173, port 49251

[ 5] local 10.105.89.112 port 5201 connected to 10.105.88.173 port 49252

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-1.00 sec 10.4 MBytes 87.3 Mbits/sec

[ 5] 1.00-2.00 sec 10.6 MBytes 89.0 Mbits/sec

[ 5] 2.00-3.00 sec 11.3 MBytes 95.1 Mbits/sec

[ 5] 3.00-4.00 sec 10.9 MBytes 91.7 Mbits/sec

[ 5] 4.00-5.00 sec 11.8 MBytes 99.2 Mbits/sec

[ 5] 5.00-6.00 sec 11.1 MBytes 92.7 Mbits/sec

[ 5] 6.00-7.00 sec 10.6 MBytes 88.8 Mbits/sec

[ 5] 7.00-8.00 sec 11.8 MBytes 99.1 Mbits/sec

[ 5] 8.00-9.00 sec 11.1 MBytes 92.9 Mbits/sec

[ 5] 9.00-10.00 sec 12.2 MBytes 102 Mbits/sec

[ 5] 10.00-10.48 sec 5.17 MBytes 91.0 Mbits/sec

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-10.48 sec 0.00 Bytes 0.00 bits/sec sender

[ 5] 0.00-10.48 sec 117 MBytes 93.7 Mbits/sec receiver

-----------------------------------------------------------

Server listening on 5201

-----------------------------------------------------------

 

What is the wifi radio chipset in your probook and 350?

Did you try multiple 350s with same results? (i.e just not 1 bad radio)

 

I think your on the right lines trying to eliminate the cause but I don't think that your get an even match between the probook and 350 due to hardware, obviously they fit better radios in probooks.

Posted
I was trying to think of the other information that was important yesterday and couldn't remember the name, today i found it MCS Index there's a good table on Demystifying Modulation and Coding Scheme (MCS) Index Values

 

This is the available speeds that wifi can connect at (spatial stream, channel band widths and Guard Intervals (wait time)). The other thing is power (which may be affected if laptop is plugged into ac or not), also if the MCS Index changes alot this can have negative impact on performance. I'd also check if you can see what the radio power levels are set. device manager -> wifi adapter -> properties -> advanced tab -> look for transmit power or power output. There's a bunch more settings there that maybe be able to be tweeked to give better performance.

 

I was interested to see here so I did a quick test between my probook (6550b) and a 215 which is a baby 350 (no 350s in my office) and both connected at 300mbs, top is probook (Broadcom 43224AG), bottom is 215 (bcm943228hmb)

 

Accepted connection from 10.105.89.10, port 2141

[ 5] local 10.105.89.112 port 5201 connected to 10.105.89.10 port 2142

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-1.00 sec 17.7 MBytes 148 Mbits/sec

[ 5] 1.00-2.00 sec 17.6 MBytes 147 Mbits/sec

[ 5] 2.00-3.00 sec 17.9 MBytes 150 Mbits/sec

[ 5] 3.00-4.00 sec 19.3 MBytes 162 Mbits/sec

[ 5] 4.00-5.00 sec 17.2 MBytes 144 Mbits/sec

[ 5] 5.00-6.00 sec 18.3 MBytes 154 Mbits/sec

[ 5] 6.00-7.00 sec 17.8 MBytes 149 Mbits/sec

[ 5] 7.00-8.00 sec 18.3 MBytes 153 Mbits/sec

[ 5] 8.00-9.00 sec 17.1 MBytes 143 Mbits/sec

[ 5] 9.00-10.00 sec 18.1 MBytes 151 Mbits/sec

[ 5] 10.00-10.03 sec 489 KBytes 174 Mbits/sec

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-10.03 sec 0.00 Bytes 0.00 bits/sec sender

[ 5] 0.00-10.03 sec 180 MBytes 150 Mbits/sec receiver

-----------------------------------------------------------

Server listening on 5201

-----------------------------------------------------------

Accepted connection from 10.105.88.173, port 49251

[ 5] local 10.105.89.112 port 5201 connected to 10.105.88.173 port 49252

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-1.00 sec 10.4 MBytes 87.3 Mbits/sec

[ 5] 1.00-2.00 sec 10.6 MBytes 89.0 Mbits/sec

[ 5] 2.00-3.00 sec 11.3 MBytes 95.1 Mbits/sec

[ 5] 3.00-4.00 sec 10.9 MBytes 91.7 Mbits/sec

[ 5] 4.00-5.00 sec 11.8 MBytes 99.2 Mbits/sec

[ 5] 5.00-6.00 sec 11.1 MBytes 92.7 Mbits/sec

[ 5] 6.00-7.00 sec 10.6 MBytes 88.8 Mbits/sec

[ 5] 7.00-8.00 sec 11.8 MBytes 99.1 Mbits/sec

[ 5] 8.00-9.00 sec 11.1 MBytes 92.9 Mbits/sec

[ 5] 9.00-10.00 sec 12.2 MBytes 102 Mbits/sec

[ 5] 10.00-10.48 sec 5.17 MBytes 91.0 Mbits/sec

- - - - - - - - - - - - - - - - - - - - - - - - -

[ ID] Interval Transfer Bandwidth

[ 5] 0.00-10.48 sec 0.00 Bytes 0.00 bits/sec sender

[ 5] 0.00-10.48 sec 117 MBytes 93.7 Mbits/sec receiver

-----------------------------------------------------------

Server listening on 5201

-----------------------------------------------------------

 

What is the wifi radio chipset in your probook and 350?

Did you try multiple 350s with same results? (i.e just not 1 bad radio)

 

I think your on the right lines trying to eliminate the cause but I don't think that your get an even match between the probook and 350 due to hardware, obviously they fit better radios in probooks.

 

Thanks for that, they both look pretty acceptable to me - would these sorts of speeds be what I'm aiming for?

 

Chipsets:

ProBook 430 G1: Mediatek MT7630E

350 G2: Realtek RTL8723BE

 

We actually haven't tried multiple 350s due to them being locked away as it's half term. On Monday I will for sure!

 

I completely agree that they aren't really a match, but I suppose at least it's discounting the Wi-Fi itself. As a test we had there today a Ubiquiti UniFi Pro and a Cisco WAP321, both requiring minimal configuration, both showing exactly the same speeds. It has to be the laptop! While I don't discredit him, as mentioned above a colleague took it home last night to try and it manages to pull 70 down from his fibre connection at home on wireless.

 

I just feel like I'm going around in circles and banging my head against a wall! :confused:

Posted (edited)

Speed is determined by MCS index, if your only connecting at 150mbs (index 7) means that your connecting on 1 spatial stream (1 radio), 40mhz channel (good), 400ns GI (good). On 1 radio that is the best speed you are going to connect at, any increase will come from multi radio wifi cards (mimo). where as my 300mbs (index 15) is same with 2 spatial streams. You could try reconfiguring your arrays to 20mhz channels and / or 800 ns GI (sometimes called long (800) and short (400)) might improve the 350 but will probably degrade probook.

 

Do an iperf against your probook and that should give a good indication of your expected throughput, there's math to work it out but i don't know what it is you should see 150mbit/s - overhead (think that's todo with frame size / MTU)

 

How far was colleague sitting from the AP (a foot :) ?) and probably only 1 or 2 wifi devices contesting air time. Worth getting him to find out what connection speed (MCS Index) he was connected at.

 

Have you tried different AP also? From the sound of the cards that its 2.4Ghz, there can be alot of other things on 2.4, wireless headphones, keyboards, mouse, speakers all can cause background interference and the probooks wifi radio power is probably a bit stronger.

 

Microwares also give off massive 2.4 interference, but normally only for a few minutes, the above wireless devicesare normally pretty constant but lower power.

 

Something else to give you some more information is something like wireshark and actually monitor the traffic to see if your getting excessive re-transmits. also looking at the switch / AP might give tx/rx errors, crc's and collisions. Collision might be the key there if probook is stronger radio other clients will hear it and shut up, lower powered might be in range of AP but not in range of each another giving cross talk.

Edited by mikeyd101
Posted

Hi, thanks for your help Mikey, will check out the MCS stuff when I can next!

 

So here is what I have come up with so far:

 

- I want to discount the laptop (350), because at home I can transfer a file at around 30MB/s over Wi-Fi. Looking online these cards are awful but it doesn't seem that this would be the case.

- I want to discount interference, although I can't prove this scientifically, I can't pick up any other wireless network in the area, and we're in a brand new building with no electric device at this point in time (the staff and students haven't moved into the building yet).

- I want to discount the wireless, because even though it's slow with my personal laptop (ProBook), it's still an awful lot better than the 350.

- I want to discount the server, because wired file transfer are like lightning (or at least in line with what you would expect from a wired gigabit network).

 

That leaves two things; the wireless controller, or the switches.

 

I am this week going back to replace one of the switches with a little unmanaged gigabit PoE switch and plug everything relevant in there. These HP switches are a little out of my depth, and haven't been configured aside from a static IP (the configuration may well be the problem, but I wouldn't know what I'm looking for).

 

The struggles goes on!

Posted

just looking at the blurb on your AP:

 

The HP MSM 802.11n Dual Radio Access Point Series works in unison with the HP controllers to deliver a high performance WLAN solution. Offering near gigabit client access and increased reliability they are ideal for voice and multi-media communications while providing full compatibility with legacy 802.11 clients and existing HP wireless controllers. Built-in Spectral Analysis and integrated Wireless IDS/IPS provides detection, classification and intelligent avoidance of non-802.11 interference and comprehensive threat protection. Access points can be deployed in managed or autonomous mode.

 

The interesting thing there is built in spectral analysis, which should mean that your able to view how busy/interference the 2.4 ghz band is (air time) and if you have AP setup on same channel giving cross talk. I'd start looking through the management interface/configuration and try and get up to speed on wifi configuration.

 

some basic wifi configuration points.

 

- 2.4ghz only really has 3 usable channels 1, 6 & 11, these are the only non overlapping channels. if your only using 2.4 don't be tempted to use any other channels it wont help.

- 5ghz is much better as every channel is usable as there's no overlapping channels

- you want APs that are in range of each other to be on different channels, or as different as possible hard todo with only 3 channels

 

- look at firmware revisions on AP / firmware revision history, look to see if there are any improvements

- I'd be contacting HP about it also see if they have any known compatibility issues

- try with as many other devices phones etc and see what kind of through put you get (is it only 350 that get bad performance)

 

Did you look at the connection speed you where connected to your home wifi?

 

From my experience most switches/AP work pretty decent with there default configuration, unless you have weird routing.

 

I hope you get to the bottom of it.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...