Exactly, remember, these are protocols desinged for the LAN environment.
Wired Ethernet uses collision detection. Wireless cannot do that because it
can't discern whether the collision is from its own source or not. So
instead, wireless takes a more novel collision avoidance approach. With
CSMA/CA, devices listen to the media befor talking. If they discover the
medium is clear, it talks. If not, it backs off for a certain contention
window. As users aggregate, contention occurs more frequently. Actual
throughput, when achieved, is no different than when the media is less full,
but we generally measure performance over time, so when more users are on
the air, each have less time to do their work.

In terms of net versus gross, the 11mbps is not myth, it is just eaten up in
the overhead of the protocol. So while the modem rate is actually 11mbps,
the total net payload for transmitted content frames is only somewhere
between 5.5mbps and 7mbps in a perfect link.

If you want to maximize throughput, engineer good links. Links that can only
sustain a connection at the 2mbps fallback rate require several x more of
the AP resources to send the same sized file as a client within the 5.5mbps
rate, etc. Few understand this, but it is the key to understanding PMP
capacity. If you are doing PTP, turn off all features required for PMP and
you will reduce the link overhead, this increasing performance of the link.

Patrick
Alvarion

-----Original Message-----
From: Sameer Verma [mailto:[EMAIL PROTECTED]
Sent: Tuesday, July 01, 2003 9:08 PM
To: S Woodside
Cc: Bulk John; 'BAWUG'
Subject: Re: [BAWUG] Throughput question - 802.11b


S Woodside wrote:

> What happens is very similar to what happens on a shared 10 Mbps 
> ethernet bus (in fact the sharing protocol is the same CSMA/CD). 

Its actually CSMA/CA (carrier sense multiple access/collision 
avoidance), which is more along the lines of 100Mbps Ethernet. The 
traffic starts backing off as it starts to get crowded, and slows down 
until things get back to normal.  Here's a nice explanation.
http://alpha.fdu.edu/~kanoksri/IEEE80211b.html.

> The maximum available rate would typically be divided by the number of 
> users. Reality is probably not quite that good. But most of the time 
> the backhaul is so much slower than the Wi-Fi bandwidth that it 
> doesn't matter.
> If you want to maximize bandwidth, use multiple radios on different 
> frequencies.
>
> simon

Sameer

-- 
Dr. Sameer Verma, Ph.D.
Asst. Professor of Information Systems
San Francisco State University
San Francisco CA 94132 USA
http://verma.sfsu.edu/ 


--
general wireless list, a bawug thing <http://www.bawug.org/>
[un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless
 
 
This mail passed through mail.alvarion.com
 
****************************************************************************
********
This footnote confirms that this email message has been scanned by
PineApp Mail-SeCure for the presence of malicious code, vandals & computer
viruses.
****************************************************************************
********
--
general wireless list, a bawug thing <http://www.bawug.org/>
[un]subscribe: http://lists.bawug.org/mailman/listinfo/wireless

Reply via email to