802.11g not a bottleneck?

I've heard people proclaim this and I've told other people this as well:

"The bandwidth of your 802.11g wireless (54 MBits/sec?) is much larger than the bandwidth of your cable/DSL (3-5 Mbits/sec), so there should be no loss of internet download speed if you switch to wireless."

However, I've found that this is not the case with my laptop. The wireless seems to be half as slow (or worse) than the wired connection. I've been running speed tests on my laptop, and I get about 5.5 MBits if I hook up an ethernet cable. I get 1.7-3.3 MBits through the wireless adapter. I know that speedtests aren't the best for getting accurate numbers, but the I'm just using it to compare my ethernet against my wireless.

I get about the same results when my laptop (an IBM X41) is 1 foot, 3 feet, or 6 feet away from my router (a Belkin). Windows' Firewall is active on both the ethernet and wireless.

What could possibly be slowing things down? Is there an easy way to gauge the wireless adapter's max speed ( just across my LAN, not across the internet )? Does the above proclamation represent the sound logic of a person who just doesn't have all the facts?

What kind of speeds to you get for internal file transfers?

Wireless is black magic dude....I had nothing but issues with it. Lots of things you cant see can figure into the S/N ratio on a wireless network...hell it could be sunspots

NetStumbler is a great tool for testing reception...it can help you get a better idea of what your signal distrobution looks like at the least

Intel wireless cards have by default a Power Saving mode enabled which messes with many routers and causes unpredictable behavior. Make sure your power is NOT set in automatic and the slider is all the way up to High. Intel hardcoded this into their drivers - the High setting is the only one that turns off their "smart bullsh*t" technology.

But the low speed of your wired connection suggests this may be a deeper problem. ALWAYS make sure your router firmware is the latest, and make sure that QoS is turned off both on the Laptop networking settings and inside the router.

Thanks guys.. I will give these a look and see.

I've been trying the speedtests today and yesterday and somehow the numbers have magically improved themselves.

Another thing is, I notice decreased wireless throughput when not using channel 1, 6, or 11.

shihonage wrote:

Another thing is, I notice decreased wireless throughput when not using channel 1, 6, or 11.

This is most likely due to channel bleed/interference and the mechanisms used to mitigate it. 1, 6, and 11 are the only frequencies that do not overlap.

Another point to keep in mind is that 54Mbit is a theoretical maximum, with most tests showing consumer WAPs hitting less than half of that in ideal conditions. Another problem with the current solutions (BGA) is that performance drops off pretty quickly when multiple devices are accessing the same WAP. The 802.11N WiFi standard (due for final approval in '07) should help resolve much of this.

Keep in mind that if you do have multiple devices accessing your router that it will default to the lowest standard of the slowest card. I.E. If you have 5 'g' devices and 1 'b', then your WAP will default to 'b' mode which will slow the connections for all.

Chum

so there should be no loss of internet download speed if you switch to wireless.

I'm glad you at least qualified that with "download speed" instead of just "speed".

One thing that's so annoying is when people treat "bandwidth" as "speed" and completely ignore that little "latency" part.

None of my consoles are hooked up wirelessly, because I don't like the added latency. My laptop is wireless, which is fine for most computing tasks, and the latency's no real issue with WoW.

Chum wrote:

Keep in mind that if you do have multiple devices accessing your router that it will default to the lowest standard of the slowest card. I.E. If you have 5 'g' devices and 1 'b', then your WAP will default to 'b' mode which will slow the connections for all.

Chum

This is the first time I hear this. Wireless "G" routers have a "mixed" mode which serves both B and G standards simultaneously. The whole point of that mode is to provide speed to the "G" devices without being bogged down by the "B" devices. Otherwise you could as well leave it in "B" mode permanently.

shihonage wrote:

This is the first time I hear this. Wireless "G" routers have a "mixed" mode which serves both B and G standards simultaneously. The whole point of that mode is to provide speed to the "G" devices without being bogged down by the "B" devices. Otherwise you could as well leave it in "B" mode permanently.

Well, I guess it doesn't technically force it into "B" mode, but it does severely cripple the throughput. For your reading pleasure. And as for leaving it in Mixed, it allows the speed to ramp back up if the "B" device is removed, whereas leaving it in "B" mode would not.

I noticed the effect when I briefly had a "B" adapter attached to my Xbox. The decrease was immediate and very noticeable. When I later removed the "B" adapter, wireless speeds returned to their normal levels.

Chum

This article mentions nothing about CTS Protection Mode which is most likely the reason for the more dramatic slowdowns. Modern routers have it disabled by default.

I believe it goes beyond CTS Protection mode and has more to do with transmission/reception mechanisms. I'll have to look around to see if I can find a good white paper on the subject.

For what's it's worth, I did some testing here in my office. Router (WRT54G) set to mixed mode with CTS disabled, all clients within 10-15 feet of AP and tested using QCheck:

Single G client - ~22.5Mb/s reported by QCheck on G client
1 G Client + 1 B Client (w/ no active traffic on B client) - ~15Mb/s reported by QCheck on G client
1 G Client + 1 B Client (B client downloading file from Microsoft.com) - ~6-8Mb/s reported by QCheck on G client

Stopping the download resulted in the G client regaining some throughput (back to 15 Mb/s), but as long as the B client was online and connected, it never returned to the baseline speed of 22.5 Mb/s. I'll grab a second G-card tomorrow and see if the degradation in speed is the same.