G devices connect when "up to 300mpbs" selected

[StratmanX]

Compared to the 4500, the 3800 I have does a better job of automatically choosing the best channel.

Remember that inSSIDer only reports on broadcasted SSID's, therefore you will not see non-broadcasted SSID's, elctrical interference, or how busy anyone's signal is concerning data throughput. All you know is that there is a signal beacon broadcasting that that person's router is set to that channel.

This is why throughtput testing becomes important to tell if the channel you have selected in less "busy" transmitting data on your neighbor's networks set at that channel as well. Throughput testing will be effected by everything that I mentioned inSSIDer does not show you.

Part of the problem with throughput testing is what looks like the best channel to select for your wireless router may become not so good if others begin to use that channel, such as after coming home from work or the kids from school. The way you'll know is experiencing slow downs in throughput. You may end up testing every channel at different times of day and night as well as on weekends when traffic can increase greatly.

IIRC, the 3800 can also change channels based on traffic, so that can help you out if a neighbor comes home and wirelessly sends a movie to his TV while you're online.

Regardless, from your posts so far, I think you will enjoy testing throughputs and using inSSIDer to help position your router and client devices to optimize wireless performance.

[Mars Mug]

Originally Posted by Zackary

I have seen iperf mentioned by many others too, so I will download that soon.

At the same time look at jPerf, a GUI front end for iPerf that can be used instead of the command line (you can still use the command line also) and which provides easier to visualise results.

iPerf will give you a much better idea of the maximum throughput you can achieve, which will be much lower than the selected wireless rate in most cases, and you might see that differences in throughput between a selected rate of 300Mbps compared to 150Mbps are not as great as expected, especially in a high contention area.

Originally Posted by StratmanX

Part of the problem with throughput testing is what looks like the best channel to select for your wireless router may become not so good if others begin to use that channel, such as after coming home from work or the kids from school. The way you'll know is experiencing slow downs in throughput. You may end up testing every channel at different times of day and night as well as on weekends when traffic can increase greatly.

I agree, there have been many threads where people have described situations where their network operated perfectly at high speed when set up, but days later there are dropouts and the rates are rubbish. Then you find through a series of questions that the wife is using a 2.4GHz video sender in the kitchen, or the next door neighbour has returned from holiday and turned on the 2.4GHz baby monitor.

As you suggest, inSSIDer may for example show three networks on channel 1, and one network on channel 6. It will not show the number of associated clients, or their network usage. Those three networks on channel 1 could be three elderly neighbours with one wireless PC each checking for e-mails and occasional web surfing, while that one network on channel 6 could be a student house with six computers using torrents and streaming videos.

Sadly, if you want optimum performance from a wireless network at all times, then you may need to monitor the environment regularly to ensure that you are using the best channel, and I’m not convinced that the automatic channel selection algorithms of most routers are robust enough to do that job well in difficult environments.

[Zackary]

Originally Posted by StratmanX

Compared to the 4500, the 3800 I have does a better job of automatically choosing the best channel.

Remember that inSSIDer only reports on broadcasted SSID's, therefore you will not see non-broadcasted SSID's, elctrical interference, or how busy anyone's signal is concerning data throughput. All you know is that there is a signal beacon broadcasting that that person's router is set to that channel.

This is why throughtput testing becomes important to tell if the channel you have selected in less "busy" transmitting data on your neighbor's networks set at that channel as well. Throughput testing will be effected by everything that I mentioned inSSIDer does not show you.

Part of the problem with throughput testing is what looks like the best channel to select for your wireless router may become not so good if others begin to use that channel, such as after coming home from work or the kids from school. The way you'll know is experiencing slow downs in throughput. You may end up testing every channel at different times of day and night as well as on weekends when traffic can increase greatly.

IIRC, the 3800 can also change channels based on traffic, so that can help you out if a neighbor comes home and wirelessly sends a movie to his TV while you're online.

Regardless, from your posts so far, I think you will enjoy testing throughputs and using inSSIDer to help position your router and client devices to optimize wireless performance.

I haven't set my 3800 to select the channel automatically, as I wasn't sure how much I could rely on it. From what you and others say though, including some reviews I've read, it sounds like it's worth giving it a go. I will (am at the moment) download iperf so I can do some testing anyway. If you're right that the 3800 can change channels based on traffic, it's definitely the option I should choose. I'll experiment with iperf and see what results I get.

InSSIDer does show one channel that doesn't have an SSID for me. There are 10 wifi routers/APs listed on there and one of them has a blank space for the SSID, but the other details still show up. It's on channel 1, like most the others.

For now I barely have any reason to need the fastest possible connection, but when funds allow I'll be buying my first NAS, so it'll then be more important. Can't wait for that.

[Zackary]

Originally Posted by Mars Mug

At the same time look at jPerf, a GUI front end for iPerf that can be used instead of the command line (you can still use the command line also) and which provides easier to visualise results.

iPerf will give you a much better idea of the maximum throughput you can achieve, which will be much lower than the selected wireless rate in most cases, and you might see that differences in throughput between a selected rate of 300Mbps compared to 150Mbps are not as great as expected, especially in a high contention area.

I'm working on that right now, and looking forward to seeing results. Right now I can't test 300mbps throughput as I don't yet have a 300mps adapter, but I'm interested to see what my 150mbps adapter gives me - Also I'd like to see how close to 1Gb my Ethernet connection gives.

[Scubbie]

Originally Posted by Zackary

...Also I'd like to see how close to 1Gb my Ethernet connection gives.

From personal experience, your actual throughput over the LAN will be determined by a number of issues, including the following:

• processor
• firewall software
• AV software
• hard drive access speed
• how busy the computer is with other tasks

I found that the latest version of ZoneAlarm crippled a couple of computers, so I removed it and found another firewall. The result was that the transfer speed doubled.

[Mars Mug]

Originally Posted by Zackary

Also I'd like to see how close to 1Gb my Ethernet connection gives.

For an unadjusted / un-optimised network you may be disappointed, I get around 300Mbps on my gigabit LAN, but it is something of a spider’s web and I have done nothing to optimise it. I think figures around 7-800Mbps are reasonable, but close to 1000Mbps is most unlikely.

By the way, I edited my post while you were posting.

Edit:

As Scubie said plus;

Cabling quality,
Interference,
Network switches,
Cable lengths,
Traffic type,
Configuration / adaptor settings,
Loads of other things I can't think of right now.

[Zackary]

Originally Posted by Scubbie

From personal experience, your actual throughput over the LAN will be determined by a number of issues, including the following:

• processor
• firewall software
• AV software
• hard drive access speed
• how busy the computer is with other tasks

I found that the latest version of ZoneAlarm crippled a couple of computers, so I removed it and found another firewall. The result was that the transfer speed doubled.

Thanks for the tips Scubbie. Whenever I do a test like this I make sure the PC is as free from other tasks as possible. I gave up with Zonealarm years ago and find the Win7 built-in firewall does a lot better job than previous versions of Windows, and actually monitors data in both directions now!