Use Multiple Channels for Faster Wireless Networking

icypyr0 writes "The Register reports: 'Current dual-mode 802.11 'a' and 'b' access points use only one of Wi-Fi's 11 RF channels at a time, with users taking turns. The Engim chipset can 'see' all 11 at once, and can use the three non-overlapping ones (1, 6 and 11) in parallel, increasing total throughput and enabling features to be incorporated in silicon that are usually implemented, at extra cost and performance degradation, in software.'"

Comments

Re:that's 110 kilometers...

[zerocool^]

I don't know about across the english channel. At some point, you pass the horizon, where you can't go any further due to the curvature of the earth. I was pretty sure that level was around ... either 50 or 100 miles, I don't remember. Can you see France from Kent?

~Will

Re:Geek factor 9.3 useful factor 1

[FatAlb3rt]

802.11g claims 54Mbps - how much faster do you need? hell, 11b is 11Mbps. that's still faster than most people's internet access.

the speed is where it needs to be (for now). i am very interested in getting a signal that doesn't crap out when i go to the other side of the house.

anyone got a link for a comparison of AP brands vs. range? comments on the linksys signal booster?

Re:Huzaa!

[Anonymous Coward]

Nope, no hope. I've setup about four dozen wireless Ethernet networks for the ISP I work for, and not a one of them works worth a damn. Well, there is one that covers only one room that works 100%, but other than that, 802.11 is complete crap. You can often get it to work for a few minutes, like these guys have, but long-term, you have non-stop problems. Even with 24 dBi antennas on each end and 100mW transmitters, the main wireless connection at the ISP I work for still doesn't work worth a damn going just across a wide street. We're using it since it's taking Hell$outh, err BellSouth, 18 months and counting to move our T1 to Sprint.

I wish a company would start making wireless that works. Instead, the marketers rule and keep pushing for bigger numbers on the boxes. Who cares if it's 5.5, 11, or 55 Mbps if it doesn't work at all?

overlap?

[StateOfTheUnion]

Are 1,6, and 11 the only channels that don't overlap?

Re:overlap?

[pe1rxq]

You could also use 2, 7 and 12 or 3, 8 and 13 as the spacing would be the same....
If you really push it you can use smaller spacings such as three or four channels instead of six.

Jeroen

Re:overlap?

[tcgroat]

This channel-hogging system is just as bad as using high-power amplifiers to blow other users off the channel. "My system is the only one that counts. I won't share with anybody else." It takes many times its fair share of a limited resource, ignoring a deliberate design strategy to use limited power (range) and reduced bandwidth. That design strategy exists to enable cooperative sharing of the available spectrum. The parallels to the me-first hooliganism on 27MHz Citizen's Band are frightening. Without cooperation and compliance to established standards and regulations, the range and through-put arms race will make wireless LANs useless to everyone.

More throughput... but...

[kneecarrot]

More throughput but more pesky interference with phones and whatnot.

Tragedy of the commons forming!

[LostCluster]

Of course, if you use the entire 2.4 GHz band, your neighbor can't. That's part of the reason why we have multiple channels to keep everybody from running into each other time. I highly doubt this group has bothered to test what kind of downside there is for a standard-issue WiFi setup operating 100 yards away.

Re:Tragedy of the commons forming!

[Anonymous Coward]

Just log in to their default-passworded APs and set them to use channel 1. Then set yours to use channel 6. Let them deal with packet loss.

As if Wi-Fi space wasn't crowded enough already...

[misleb]

I've been in places where you can't find a free channel because there are too many other people using Wi-Fi in the area. Now we're going to have individuals using ALL of the non-overlapping frequencies? That really sucks. As far as I am concerned, Wi-Fi as we know it will not go very far. There is just no way people are going to be able to share such a limited frequency spectrum... not in densly populated areas, anyway. I'll take good ol' reliable wires any day. Wireless is overrated.

-matthew

Interference

[Ctrl-Z]

This type of idea is not new, and I have seen it in wireless routers/nics for months. The primary drawback is that if you are using up all those channels, your neighbour's wireless network won't have anywhere to go. Conversely, if you are that neighbour, it wouldn't make you very happy.

I consciously decided against buying something like this for that very reason when I bought my wireless hardware, even though the cost difference was negligible.

Re:Use Multiple Channels

[Anonymous Coward]

Yeah, instead of driving a long vehicle with multiple rows, everyone could be in the front seat! Too bad if it makes it difficult for others to change lanes or bypass problems.

Valid Uses

[yoshi_mon]

While it's been duely noted that using up more channels could interfere with others who are trying to use 2.4, there are a number of applications I can see that would be useful for this type of setup.

Large old office buildings that arn't wired for ethernet, large warehouses, and people who live on large plots of land.

Yes, if you are living in a typical burb or in the city and try to use one of these you could run into issues with running out of channels. However not everyone lives/works in small areas.

Re:As if Wi-Fi space wasn't crowded enough already

[LostCluster]

Yep... but you have to be careful what you're using is right for your application. If you're using WiFi to go accross the room, you should be using Bluetooth.

Sounds like

[ajs318]

This just sounds like a new form of bogarting to me. The whole idea of there ever being multiple channels in the first place was so that everybody would have a chance of finding a free one -- things like this go right against that idea.

If it isn't actually illegal, it's certainly anti-social. But then again, I don't use any wireless kit anyway ..... I need to have a power cable, so I don't mind having a network cable as well.

Huh?

[blair1q]

If I'm buying a new chipset, I'm buying 802.11g and getting 5X the speed, not just 3X.

Re:What about 802.11G?

[timeOday]

If I understand correctly, you're saying 10base2 is shared while 10baseT is not? That's not true. Both wire types are used for ethernet, and both are shared. Ether way, collisions aren't that big a deal after all, see the classic reference:

Ethernet works in practice, but allegedly not in theory: some people have sufficiently misunderstood the existing studies of Ethernet performance so as to create a surprisingly resilient mythology. One myth is that an Ethernet is saturated at an offered load of 37%; this is an incorrect reading of the theoretical studies, and is easily disproved in practice. This paper is an attempt to dispel such myths.

...

Figure 10 shows excess delay , a direct measure of inefficiency. It is derived from the delays plotted in figure 8. The ideal time to send one packet and wait for each other host to send one packet is subtracted from the measured time. The time that remains was lost participating in collisions. Notice that it increases linearly with increasing number of hosts (offered load). When 24 hosts each send 1536-byte packets, it takes about 31 milliseconds for each host to send one packet. Theoretically it should take about 30 mSec; the other 1 mSec (about 3%) is collision overhead. Figure 4 agrees, showing a measured efficiency of about 97% for 1536-byte packets and 24 hosts.

The upshot is that ethernet can carry very close to its rated capacity even if there are a lot of hosts and a lot of collisions. (Of course nowadays we tend to use switches instead of hubs anyhow, but that's not a at all inherent in 10baseT wiring).

Whether wireless will work quite this well, I don't know.

Re:What about 802.11G?

[42forty-two42]

Oh, and 1000baseTX is a switched-only Ethernet(Of course, it's only really ethernet due to marketing decisions).

Actually, it's called "ethernet" because its frame format is identical to that of the slower variants, and it supports autonegotiation to be backwards-compatible with 10Base-T and 100Base-TX.

Re:tradeoff

[wankledot]

50 is a great number to design around, because any more than that you will be cut down to a trickle of bandwidth for each one. The access points technically support quite a few more simultanous connections than that, but no one is going to suggest that you design a newwork for a thousand clients per AP. There might be some odd situation where you need more than a hundred clients per AP for some low-bandwidth use (can't think of one, maybe like, a zillion iPaqs in a big room with one AP.. who knows.

IOW, you could have a lot more than 50, but you really shouldn't.

Noooooooooo!

[Jahf]

This will destroy wireless ISP communities.

I already have only -2- channels that I can reliably use in my house without interference. Every other channel is in use for ISP access in our community or gets interference from cordless phones and microwaves.

If you want more throughput, use different frequencies. Even if they are close to 802.11b/g that is better than going into the already established spectrum.

Yes, I know that this is not mandated or regulated space, so there is not much I can do to enforce my needs. However unregulated waves only work if people make an effort to play well together.

Re:Some comparrisons regarding G

[rfmobile]

I will place 2 or 3 AP's on different channels with the power turned all of the way down so it will balance the user load between the them.

This is a great example of an intelligent use of the 802.11X protocols. Use only as much power as you need, and only where needed. As opposed to the usual (and wrong) tactic of using full-power in an attempt to squash a neighboring access point's signal.

-rick

Re:Department

[BandwidthHog]

Well, the post was the post, the sig is just a sig. But if you'd like I'll gladly explain it. It is meant to remind people that you shouldn't do things that adversely impact the lives of others based on your superstitions. At least that's my take on it, I didn't come up with it (although I wish I had).