cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
1211
Views
5
Helpful
15
Replies

Trying to understand Spectrum/Channel and Bandwidth..

yprasannas
Level 1
Level 1

In a hall lets say for a conference we install 4 AP's on each corner assuming that this will elevate the bandwdith issues when streaming the video.

Does this mean that 4 AP's will give 54Mbps of Bandwidth so the total throughput is 216Mbps?

One guy is saying because they all close enough that 2.4Ghz spectrum is shared by all so by the end throughput will only be 54Mbps.

Appreciate if somebody can give overview of the spectrum,bandwidth?

Also for an AP there r 11 channels. Does each channel has 54 Mbps or 11 channels put together in 2.4Ghz spectrum is total 54Mpbs?

15 Replies 15

sschmidt
Cisco Employee
Cisco Employee

It would be 54 Mbps total. Take a look at the Wikipedia entry for 802.11 here:

http://en.wikipedia.org/wiki/802.11

Lots of good information to help clarify everything.

802.11 g has only 3 no over lapping channels i.e 1, 6 and 11.

Does it mean that if u fire up 3 AP's close to each other then each one will choose each channel among 1,6 and 11?

In that case why can't each AP has throughput of 54 Mbps and 3 AP's willl have 3*54 Mbps throughput?

Thanks for your time,

You could set the AP to use the channels, so there is no overlap, but this does not mean that the throughput is going to go up. Potentially, you could have 1 client attach to each of the AP;s, and each one could get 54Mbps. Remember that wireless is half-duplex, can only listen or receive to one client at a time.

HTH,
Steve

------------------------------------------------------------------------------------------------
Please remember to rate useful posts, and mark questions as answered

So lets say 3 AP's get 1 channel each and 1 user attaches to 1 AP each. They are all receiving a media stream. So throughput is 3*54 Mbps(theoretical). is that right?

In that case installing more than 3 AP's does not help in a closed area because the AP's will start sharing channels and the throughput gets shared.

Each host/NIC can only associate with one AP at a time.

Also understand that 54Mbps is the signaling rate, not the throughput. With a perfect connection at an ideal distance, on a host that is capable, with properly written drivers, running all by itself (i.e., not sharing resources) ... you can get a bit over 22Mbps to that one host.

If you add hosts, that 22Mbps is shared / split according to the application requests.

If the signal is degraded in any way (noise, low signal quality, some human stands between the host and the associated AP), you will NOT get ~22Mbps, you'll get significantly less.

There is no sane way to consolidate bandwidth with 802.11g or 802.11a. You might be interested in 802.11n.

IMO, you should call a professional and have a eye-to-eye discussion and comprehensive site survey. It sounds like you are making plans on some seriously flawed assumptions, and are willing to be way too optimistic and assume the best or a perfect situation ... when in reality, with wireless, it is almost always the worst of the worst, the situations cascade against you, and both Murphy and Darwin both hate you.

Always plan for the worst, you won't be disappointed. One of my favorite sayings is "Optimists are never pleasantly surprised" ... if you want to design networks of any type, you have to be pessimistic, honest, you do ... ask anyone that's been doing them for a while. Most can count 100% success on the first effort on one hand or less (i.e., "No battle plan ever survives first contact with the enemy")

Good Luck

Scott the Pessimist, Curmudgeon at Large

Yes I understand the 22MBps.

My question was more generic on trying to understand the concept of channels/bandwidth/radios.

Do u have any input on these generic concepts?

Thanks!

Well, there's generic, and then there's generic ... it can be as simple as thinking of Television broadcasts:

* There is a large chunk of spectrum carved out for TV

* To organize that spectrum practically and administratively, it is subdivided into channels

* To get programming located within one of those channels, you tune to the centerpoint of the channel.

Same thing for AM, FM, and Shortwave radios ... like TV, each channel is assigned assigned a specific amount of bandwidth (a portion of the band, which is a portion of the spectrum of DC-->cosmic rays).

802.11{b, g, a, n) is no different. They operate within a given band, on specific channels (portions of the band).

There are other differences, most significant being the modulation (how the data is imprinted onto the carrier), and the methods used to improve the chances of getting the data all the way through to the destination error-free (or at least in a condition where the errors can be corrected, instead of having to re-transmit the traffic).

What about the radios don't you understand?

What *IS* your understanding of channels, and bandwidth? Tell us what you think it is, and if it's not right, we can provide some guidance.

We need a starting point, because there's not enough room on the forums to give you a complete background in RF theory, even if I/we was smart enough to deliver it.

Summary:

BAND = administratively assigned portion of some RF spectrum

CHANNEL = Administratively assigned segment of the band

BANDWIDTH = the measure of "How Much" spectrum is under discussion

THROUGHPUT = How much data you can pass via the assigned bandwidth, this can vary with the type of modulation and, multiple modulation methods can be used within a specific band or channel ... but they are usually not compatible and can't "see" each other.

RADIO = a device to receive and demodulate (separate) the data/traffic from the carrier.

Scott

Scott - or anyone

Is there a way of creating a theoretical model for the following to derive expected throughput (goodput) or is it better to practically measure it. (I may need to borrow some laptops :-) )

All based on Single 802.11g AP and

10 x 802.11g client. no co-channel or

adjacent channel interference

1) 802.11g AP data rate 54mbps

no encryption

Theoretical Throughput 22mbps

due to (CSMA/CA)

10 x 802.11g wireless clients

All clients transmitting this data

simultaneously

Data http (TCP) <-(will add overhead)

Packet size 1460Bytes

Latency incurred between wireless clients

and http server 10ms <-(will add overhead)

2) 802.11g AP data rate 54mbps

WPA (TKIP) encryption - (will add overhead)

Theoretical Throughput 22mbps

10 x 802.11g wireless clients

All clients transmitting this data

Simultaneously

Data http (TCP)

Packet size 1460Bytes

Latency incurred between wireless clients

And http server 10ms

Latency incurred between wireless clients

and http server 10ms

3) 802.11g AP data rate 54mbps

WPA2 AES encryption

Theoretical Throughput 22mbps

10 x 802.11g wireless clients

All clients transmitting this data

Simultaneously

Data http (TCP)

Packet size 1460Bytes

Latency incurred between wireless clients

and http server 10ms

(In my opinion ...)

Unless it is a static, unpopulated, empty room of a known exact dimension, with all construction materials and surfaces known, and all possible interfering sources are known, measured, and static ...

No, not for the purposes of design. The "heat maps" used with wireless management work because they're taking live measurements.

To try to anticipate, then code, all the possible variables for the sake of engineering a design would be a monumental task.

It's easier and much more reliable to eyeball the environment and make some tests with active equipment.

Good Luck

Scott

Scott

I have found this document detailing 802.11b

tests

http://www.cisco.com/en/US/docs/wireless/technology/channel/deployment/guide/Channel.html#wp134413

Do you know of one detailing 802.11g

Mark

That document also details 802.11g (for the topic it's discussing).

The short story for that doc is that while you *might* be able to slide in a four-channel 802.11b install, it would be a Bad Thing to try it with 802.11g because the energy distribution in an OFDM signal is still too high at the 11Mhz shoulder (-20 vs -35).

Why not just set up an AP, and do your own walk around? It's still invalid for the purposes of design and specification, but a lame excuse ("according to our office tests ...")is better than none. Any customer that's been exposed to wireless principles will recognize it as bogus.

Documentation will not help here; it is purely an environmental issue specific to that exact location. You should do a survey. If you're NOT willing to do a survey (or the customer refuses to allow you), then drop it; walk away ... you're setting yourself up for a long stretch of gut-wrenching nightmare.

Good Luck

Scott

Scott

Thanks

Mark

Mark/Scott,

Thanks for posting that link. I had not run across it before.

Great discussion.

-chris

ScottMac,

Excellent stuff mate. A 5 from me for sure.

Ken

Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: