[gentoo-user] WLAN performance

[gentoo-user] WLAN performance

[Florian Philipp]

[-- Attachment #1: Type: text/plain, Size: 318 bytes --]

Hi list!

Can anyone tell me how I could find the bottleneck on my wireless
network? According to iwconfig, both cards are cofigured for 54 MBit/s
but I only get about 13 MBit/s, both on NFS and scp.

Might it be the driver (iwl3945, rt61 from kernel 2.6.24)?
Unfortunately, iwspy doesn't work on these cards.

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-user] WLAN performance

[William Kenworthy]

ahh, marketing.  Some people will believe anything!

Check the table at
http://www.oreillynet.com/pub/a/wireless/2003/08/08/wireless_throughput.html

or numerous other guides courtesy of google.  This shows that maximum
throughput is roughly 27Mb/s for 54g, but in my experience its much less
in the real world.

Note that configuring a card for 54mb/s is the maximum - unless you are
quite close (distance wise), have little interference and dont have a
busy 802.11b on the same AP, you are not even going to see a 54, but a
fallback.  And I think encryption will clip it even further if you are
using that (as you should be!)

BillK


On Fri, 2008-03-21 at 11:33 +0100, Florian Philipp wrote:
> Hi list!
> 
> Can anyone tell me how I could find the bottleneck on my wireless
> network? According to iwconfig, both cards are cofigured for 54 MBit/s
> but I only get about 13 MBit/s, both on NFS and scp.
> 
> Might it be the driver (iwl3945, rt61 from kernel 2.6.24)?
> Unfortunately, iwspy doesn't work on these cards.
-- 
William Kenworthy <billk@iinet.net.au>
Home in Perth!
-- 
gentoo-user@lists.gentoo.org mailing list

Re: [gentoo-user] WLAN performance

[Stroller]

On 21 Mar 2008, at 10:33, Florian Philipp wrote:
> ... both cards are cofigured for 54 MBit/s
> ... I only get about 13 MBit/s, both on NFS and scp.

I'd be really quite happy with that.

As BillK remarks, 50% of your 54 Mbit is consumed by protocol  
overhead. You're probably going to tell us that the two machines are  
currently right next to each other, so losing an additional 50% to  
interference & other intangibles might seem at first sight  
unreasonable, but I doubt you'll ever do any better than that, and  
you could waste a lot of time trying.

Wireless is for surfing on the sofa - you can easily get eighty times  
this throughput with a cable, so investing time & energy in trying to  
get a two-times performance increase is a poor return.

Stroller.
-- 
gentoo-user@lists.gentoo.org mailing list