Powerline/EOP wifi in house ~roaming~

Forums › LAN (ethernet/Wifi/routers/Bluetooth) › Powerline/EOP wifi in house ~roaming~

1101

3122 posts

Uber Geek

#191452 4-Feb-2016 16:16

Ive a question re Powerline/EOP wifi kits

Ive setup the remote powerline device to have a different SSID than the router's Wifi, to ensure that the laptop/whatever will be
connected to the most appropiate wifi access point . ie if downstairs, the manually connect laptops wifi to "downstairs" SSID & vice versa when upstairs

If I set the powerlines SSID to match the main wifi AP's SSID , so there is just the one wifi SSID, I assume I may run into issues with devices not connecting to the strongest signal AP ? ie when downstairs the laptop may try & connect to the upstairs access point/router

Jase2985

13466 posts

Uber Geek

ID Verified

Lifetime subscriber

#1485324 4-Feb-2016 16:33

correct, the laptops wireless adapter might be able to change the roaming aggressiveness.

you could try a program like NetSetMan and configure it to auto switch based on the signal strength. ive found its good at switching based on SSID but havent tried based on signal stregnth.

hio77

12999 posts

Uber Geek

ID Verified

Trusted

Lizard Networks

#1485383 4-Feb-2016 17:49

make sure to keep the authentication the same, Different channels - Preferably non overlapping.

I have not come across a device in the house yet, that doesnt roam along some are slower than others to realize its time to hop to another AP though.

#include <std_disclaimer>

Any comments made are personal opinion and do not reflect directly on the position my current or past employers may have.

Crowdie

228 posts

Master Geek

#1495217 18-Feb-2016 20:24

There are three key areas to look at to improve your wireless client roaming performance:

Transmit Power. What transmit power do you have the 2.4 and/or 5 GHz radios configured at? If you are going to support tablets and smartphones you don't want to have a transmit power greater than 15 dBm. Finding the "correct" transmit power for your house will require a bit of "trial and error" until you get the 2.4 and 5 GHz transmit powers so your wireless clients seamlessly roam between the access points/wireless routers.
Data Rates. Do you have the low data rates enabled? If you do your wireless clients will become "sticky" and be reluctant to roam. As a rule you disable the 802.11b data rates (1, 2, 5.5 and 11 Mbps) to eliminate the 802.11g protection mechanisms overhead (unless you have 802.11b only wireless clients you need to support) and I commonly disable the 6 and 9 Mbps data rates so the lowest supported data rate is 12 Mbps.
Antenna Selection. You are most likely using wireless routers/access points with omni-directional antennas. If you have a wireless router with standard 2.14 dBi dipole antennas then the signal propagation through the floor/ceiling will be higher and wireless clients will become "sticky". A higher dBi omni-directional antenna (5 or 8 dBi, for example) will "flatten" the signal propagation so more signal will be directed across the floor and less through the floor/ceiling. If your access point/wireless router has internal antennas then you cannot change them.

If your wireless devices do not, at least, allow you to adjust the transmit powers and data rates then you are going to find it difficult to optimize your client's roaming performance.

raytaylor

4017 posts

Uber Geek

Trusted

#1495330 18-Feb-2016 23:35

Its always best to use the same SSID, passkey type and passkey, but different channels.

The roaming stickyness can be controlled on some laptops, but not on tablets. But you can turn down the power level of the APs to force the device to roam sooner.

Eg. if there is a power output setting, run the wifi access points at 70% tx power.

You will have roaming capability with smaller "cells"

Ray Taylor

There is no place like localhost

Spreadsheet for Comparing Electricity Plans Here

Crowdie

228 posts

Master Geek

#1495751 19-Feb-2016 15:53

The issue with only using transmit power changes to improve roaming performance is that you can introduce asymmetric power issues, which will increase the number of retries and, hence, increase the latency/jitter and reduce the throughput. Unfortunately a large number of consumer wireless routers do not allow you to change the data rate settings so you don't really have a choice. If the consumer wireless router has the 802.11b data rates set to mandatory (these settings may not be accessible via the router's GUI and can only be seen in the beacon frames) then it is going to be extremely difficult to get the 2.4 GHz wireless clients to roam in a small two level dwelling, for example.

As more people start using latency/jitter sensitive applications, like VoIP and Netflix, over their wireless networks these types of issues will become more obvious. I have seen a number of posts saying that wireless isn't any good but when done correctly it is very good. The issue is that to configure a wireless network correctly:

You need to understand the 802.11 protocol. The CWNP program is exceptional at teaching this.
You need to have the right equipment. A professional site survey application (AirMagnet Site Survey Pro or Ekahau Site Survey Professional, for example), spectrum analyzer (AirMagnet SpectrumXT or Metageek Chanalyzer, for example) and protocol analyzer (AirMagnet WiFi Analyzer or Metageek Eye P.A. for example) are the starting points. For residential use the free Ekahau HeatMapper application may be suitable for site surveys.
You need to have experience working with wireless networks.

Unfortunately, we have far too many people deploying wireless networks in New Zealand that just don't know what they are doing.

1101

3122 posts

Uber Geek

#1495817 19-Feb-2016 17:01

Crowdie:

Unfortunately, we have far too many people deploying wireless networks in New Zealand that just don't know what they are doing.

???

This topic is home wifi, over powerline.

Are you seriously going to charge the time & $$$ ,to do site surveys, spectrum analysis etc just to setup a $99 wifi printer or give a kid
wifi access in his room ?
Thats the reality

The EOP unit I installed last week, the software had no options to make any changes to the EOP wifi unit. The option was push the clone button, thats it :-(
Luckily , it works , so far.

Crowdie

228 posts

Master Geek

#1495825 19-Feb-2016 17:40

1101:

Are you seriously going to charge the time & $$$ ,to do site surveys, spectrum analysis etc just to setup a $99 wifi printer or give a kid
wifi access in his room ?
Thats the reality

I site surveyed my neighbour's house (the spectrum analysis is done at the same time) and I was out of there in thirty minutes. Now most people don't have the equipment to do this but they then have to accept that their performance may not be that great - which is more what the comment was about. Some of the consumer grade wireless products are really, really poor - it is almost impossible to get good performance out of them. When are are WiFi Alliance certified they only get the minimum mandatory features tested so they can use the WiFi logo. It is the optional features that make wireless work well and these are make the more expensive wireless products out perform the consumer grade products. So the issue is not the wireless technology but the consumer grade products selected.

If you are just going to web surf and watch YouTube videos then you are not going to notice the difference but, as I said, more and more people are streaming videos and/or deploying VoIP and any money you save with the consumer grade products you will lose twenty times over in frustration.

When you are deploying wireless in a residential site try the following:

Avoid deploying latency/jitter sensitive applications in the 2.4 GHz spectrum. This may require you to have multiple SSIDs - one in 2.4 GHz and one in 5 GHz.
Only use 20 MHz channels in the 2.4 GHz spectrum.
If you are having coverage issues with the 5 GHz channels drop the channel width to 20 MHz. This will raise the RSSI (signal strength) by 3 dB (assuming you were originally using 40 MHz channels) - basically it will double the signal strength.
Only use the 5 GHz non-DFS channels (36, 40, 44, 48, 149, 153, 157, 161). Channel 165 is non-DFS but a number of wireless devices don't support it so it is best to avoid it.
If you can disable the low data rates do.

Quic Broadband

Move to New Zealand's best fibre broadband service (affiliate link). Free setup code: R587125ERQ6VE. Note that to use Quic Broadband you must be comfortable with configuring your own router.

1101

3122 posts

Uber Geek

#1496940 22-Feb-2016 13:11

Crowdie:

I site surveyed my neighbour's house (the spectrum analysis is done at the same time) and I was out of there in thirty minutes.

A 30 minute site survey & I assume a real/total 2.4Ghz analysis & not just wifi , will only give results for that 30minutes .
It is therefore of limited use as other sources of interference startup & stop during the day & night.
When I rang a specialist and asked about having site survey done, he said it was a waste of time as it only shows issues at the actual time of the Survey

case in point: a client with wifi that would cut out when her neighbors came home at night. Obviously a site survey wouldnt detect or help there unless
you ran it all week.

The real fix for 2.4 is not to use it, run cables

Im guessing your work is more at the Enterprise Level , rather than the home user who wants a $50 fix for their wifi thats not usuable on the 3rd story of the house
Cheers

Crowdie

228 posts

Master Geek

#1497077 22-Feb-2016 16:05

The site survey/spectrum analysis gives me a list of interfering devices (microwave ovens, cordless phones, cell sites, etc.) and which part of the building there are affecting. Without knowing the environmental conditions you haven't got a chance of resolving the issue. An issue we are starting to see is raised 2.4 GHz noise floors caused by the new cell sites being deployed. A user wouldn't be able to detect the raised noise floor, as they don't have spectrum analyzers, but will experience the random disconnections.

Even if a neighbour had a poorly configured wireless network they shouldn't generate enough duty cycles to break a wireless network. When you do the site survey if you see large numbers of poorly configured wireless networks, which is not uncommon, then you know that if the majority of them were in active use then enough duty cycles could be consumed to cause a problem.

I agree with you that in residential and CBD areas the 2.4 GHz spectrum has become extremely crowded and latency sensitive applications will perform extremely poorly.

The point I am trying to get across is that low end consumer devices are not suitable for anything more than casual web surfing. The applications that people are now wanting to run (video conferencing, HD video streaming, etc.) need better quality wireless equipment and a $20 wireless card talking to a $100 wireless router is just not good enough anymore.

There was a really good white paper written in the UK around eighteen months ago that said the consumer wireless vendors were one of the biggest risks to the wireless industry. They produce poor quality products and use a combination of ambiguous specifications/marketing combined with general end user's general lack of knowledge about wireless to push their products.