linux-wireless.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* Implementing wireless testbed on Linux
@ 2009-08-21  8:26 Jinsung Lee
  2009-08-24  8:18 ` Holger Schurig
  0 siblings, 1 reply; 2+ messages in thread
From: Jinsung Lee @ 2009-08-21  8:26 UTC (permalink / raw)
  To: linux-wireless

Hello, everyone.

I'm a Ph.D. student and I'm going to implement wireless testebed consisting
of more than 20 nodes using WLAN device on Linux.
Actually, there are a lot of devices and drivers in the world. I can't
follow all information as you know, which is very difficult job. 
So, I expect to borrow very valuable knowledge and experience from you.
Specifically, what I want to do is to set 802.11 MAC parameters such as
AIFS, CW, TxOp in per-packet basis in user-space.
In the case of using 802.11n chipset, which card and device driver should I
choose? At the same time, which version of Linux kernel should be
appropriate?
Is there any difference between pci and usb interfaces?
Definitely, all device and software should be stable and up-to-date.
Please let me know anything you know related to this issue.
I'm looking forward to your comments.
Thanks.

----
Jinsung Lee



^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: Implementing wireless testbed on Linux
  2009-08-21  8:26 Implementing wireless testbed on Linux Jinsung Lee
@ 2009-08-24  8:18 ` Holger Schurig
  0 siblings, 0 replies; 2+ messages in thread
From: Holger Schurig @ 2009-08-24  8:18 UTC (permalink / raw)
  To: Jinsung Lee; +Cc: linux-wireless

> In the case of using 802.11n chipset, which card and device
> driver should I choose? At the same time, which version of
> Linux kernel should be appropriate?

I'd go with Atheros based PCI cards. The cards are good, and the 
support is outstanding (Atheros people work together with Linux 
community on this mailing list!).

> Is there any difference between pci and usb interfaces?

Sure. Please read http://www.pcisig.com/specifications and 
http://www.usb.org/developers/docs/. The busses are VERY 
different in case of hierarchy, latency, max transfer rates and 
so on.

I fear, however, that this mailing list isn't the right channel 
to transfer this knowledge.



> Definitely, all device and software should be stable and
> up-to-date.

That's a contradicion. Especially wireless is quite evolving in 
the Linux kernel.


There's the possibility that you use a vendor-kernel (e.g. Debian 
kernel, Fedora kernel, or whatever distro you use). I personally 
won't do this, but I'd use a more "vanilla" kernel. The vendors 
usually put a huge amount on patches into their kernels, that 
will make results not as easy to compare. Also, if you have a 
specific questions, chances are way higher that the people in 
the mailing list know about the "vanilla" kernel or 
wireless-testing kernel than some random vendor kernel.


If you absolutely need something stable, you should go with the 
least recent stable linux kernel (see http://www.kernel.org/).

If you want to be more up-to-date, you should use 
wireless-testing. Expect new functionality, fixed bugs, but also 
new bugs there :-)    However, If you ever have to change some 
driver code, you definitely want wireless-testing. You cannot 
(usually) use the source from the stable kernel when submitting 
patches. More info at 
http://wireless.kernel.org/en/developers/Documentation

^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2009-08-24  8:18 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2009-08-21  8:26 Implementing wireless testbed on Linux Jinsung Lee
2009-08-24  8:18 ` Holger Schurig

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).