Wireless has come in in leaps and bounds over the years. While it’s freed many people from Ethernet cables, it’s also more complex and error prone than a cabled network. The network industry has also been ruthlessly ‘efficient’ in its marketing with regards to speed and performance claims, this guide sets out some of the basics when troubleshooting or performance testing, and will assist in resolving the vast majority of issues.
Speed and reliability issues generally come down to one of four things:
1. Incorrect expectations
4. Mixed Mode settings.
Get your expectations right.
Speed measurement: Wireless access points or routers are measure in megabits per second (Mbps). Do not confuse this with Megabytes (MBps) per second, which you may see when transferring files in windows computers. 1 Megabyte = 8 Megabit, so if you’re measuring speed by copying large files in windows, multiply it by 8 to get to your correct speed. Speed should be more or less synchronous – ie, the same in both directions.
Measure at the correct place: Many people go to sites like Speedtest.net to measure speed. However, while this can be useful, if your wireless speed is faster than your internet speed (which is likely), you are actually measuring your internet connection speed, not your wireless. While there is software to do this accurately, generally finding a 100 megabyte file and copying between windows machines suffices for basic performance testing. This also means that if you use your wireless just for internet access, there’s no point spending lots of money on a very fast AP if you have a relatively slow internet connection.
Lies, damn lies and 802.11 statistics.
When promoting their Access points, manufacturers use ‘theoretical maximum’ figures. However these figures are so theoretical, that they can never actually be achieved for file copying, even in a test lab. So when troubleshooting or testing, the first step is to ignore any speed figures on the box.
The following table sets out claimed speeds, real world speeds in good conditions and time to transfer a 100 Megabyte file, which is a handy way to see if the network is performing as expected. If you’re consistently within 25% either way, the network is fine, and the problem is expectations set by unrealistic claims. If not, proceed to step 2.
Claimed Real 100MB file transfer
802.11b 11mbps 4-6 Mbps ~2.5 minutes
802.11g 54mbps 18-22 Mbps ~40 seconds
802.11n 300mbps 40-50 Mbps ~20 seconds
*This ignores any channel bonding or manufacturer specific improvements, which some AP’s may implement. These muddy the water a bit more as not all clients can take advantage of this.
2. Interference & channels
The 2.4ghz spectrum can be pretty congested. It consists of 13 channels which overlap by a couple of channels either side. So an Access point using channel 3 will slow an access point competing for the same bandwidth using channel 1,2,4 or 5. A simple tool such as netstumbler will allow you to see what other AP’s nearby are using – this is a great tool sepecially if provisioning in a ‘crowded’ area – ie a shared office block or high density housing. Bear in mind that this will only show you other wireless networks – not general interference, but it’s often all that is needed.
In short, pick the least utilised channel you can, with little or no overlap if at all possible. Many AP’s will autoselect a channel, but often they don’t do this particularly efficiently, so change the channel to one less congested.
Some AP’s and clients can operate in the 5ghz range – this may be worth a try if the 2.4ghz range is very congested, however this is also likely to reduce coverage and penetration through walls etc, so best used in an open plan type area.
If using multiple AP’s for roaming, use the same SSID and key, and put them on different channels – for example if you had 3, you might use 1, 6 & 11.
Other AP’s aren’t the only thing to cause intererence – many electrical devices can – microwave ovens, phones, radios. Try and move these away from the problem area if possible. If the problem is intermittent, look for devices that don’t work continuously. A Wi-Fi Spectrum analysers may be useful in such situations and aren’t hugely expensive for a USB based one.
Some good further reading here
Wireless doesnt like big thick walls much. It can also get confused by large metal items nearby. Try experimenting with the placement of the access point to find the sweet spot offering the most coverage. This is easily tested by placing the AP, and having a wander around with your laptop while keeping an eye on the amount of bars you have. Ceiling mounting or wall mounting the AP may improve things dependent on the building and where access if required.
4. Mixed Mode settings.
Most access points will slow down significantly when in mixed mode – ie when supporting both B & G or G & N etc. Set this to ‘N only’ or ‘G only’ if all your devices support it. Most devices made in the last few years will support N, which is preferable.
RSCC Support both wired and wireless LANS in Hitchin, Stevenage and Letchworth, please call on 01462 458571 should you need any assistance with yours.