Re: [PLUG] non-standard 11g

[Jason Costomiris <jcostom@jasons.org>]

On Jun 17, 2004, at 8:18 AM, Tobias DiPasquale wrote:

> On Thursday 17 June 2004 07:53, Jason Costomiris wrote:
> | The Linksys stuff does it by waiting to transmit packets.  In other
> | words, buffer the smaller packets to be sent over the air and instead
> | send bigger chunks of data.  The result is a very slight increase in
> | latency (due to the bufferring), but a 30 - 50% increase in available
> | bandwidth.  Not bad.
>
> Where did you hear that? I just bought a WRT54G and eschewed the one 
> with
> "SpeedBooster" as being nothing but some software hack for $50 more. 
> If what
> you're saying is true, then I was correct. However, I told some guys 
> at the
> office about it and one of them said that Linksys was also using the
> dual-channel approach in order to reach 108Mbps with the SpeedBooster 
> WRT54G
> and that the SpeedBooster and non-SpeedBooster varieties were actually
> different hardware.
>
> Do you have any links detailing what the difference really is?

Links?  Nope. :(

It's in an article found in the July 2004 issue of MobilePC Magazine 
(pg. 12 if you have a copy).

The article states, "Speed-boosted Wireless G routers promis as much as 
108 Mbps of throughput, twice as fast as standard Wireless G products.  
How do they do it?  First, through packet bursting, in which shorter 
data packets are rebuilt into longer, more efficient ones that transfer 
data more quickly.  Secondly, these products compress data whenever 
possible. Broadcom 108 Mbps chipsets, found in Linksys's Speed booster 
108 Mbps routers and adapters, use both of these methods to speed up 
transmission."

It goes on to say, "Atheros chipsets, found in Netgear and D-Link turbo 
products, also use a third, more controversial, speed trick called 
channel bonding.  Instead of using one channel to transmit data -- the 
traditional WiFi way -- these products use two channels simultaneously 
to send more data.  But in taking up more space on the channel 
spectrum, you may be leaving less room for other wireless networks near 
you."

The conclusion about the slight added latency was my own.  Stands to 
reason...  You take small packet data, wait for more of it to show up 
in the router, then re-pack the data into larger packets before 
transmission.  Clearly that's additional latency, right there.

--j

___________________________________________________________________________
Philadelphia Linux Users Group         --        http://www.phillylinux.org
Announcements - http://lists.phillylinux.org/mailman/listinfo/plug-announce
General Discussion  --   http://lists.phillylinux.org/mailman/listinfo/plug