[Olsr-dev] does the current Debian olsrd.conf work with a mesh?
Hans-Christoph Steiner
(spam-protected)
Fri Jun 29 20:29:46 CEST 2012
On Jun 29, 2012, at 1:49 PM, Markus Kittenberger wrote:
> On Fri, Jun 29, 2012 at 7:06 PM, Hans-Christoph Steiner <(spam-protected)> wrote:
>
> IMHO, why would I blindly follow rules when I have actual, real world
> data that clearly demonstrates the problems with those rules?
> which are?
>
> That wikipedia page shows the theory quite nicely, but it does not take
> into account so many real world factors.
> which are?
>
> likely that too many people that think they are smarter than all others?
> It shows a picture of a clean
> room lab environment,
> read it again,..
>
> no this is no lab setup it just a chart showing the idea behind recommending this channels.
>
> which btw are the optimum usage of the band!
>
> and any user that uses another channel reduces the overall throughput in the band!
>
> and as your scans showed that most networks in your region are actually on 1,6,11, the only ermaining thing to do, is choose the least used out of the 3.
>
> and even if u measure e.g. real world data throughput on each channel, and find e.g. 8 to be the best, you (in 99 of 100 cases) still should not use any channel except 1, 6, 11 (-;
>
> and especially not as a default for others!
>
> where you have complete control over all of the
> radio waves in the 2.4GHz band.
>
> Working with radio waves means understanding the theory, but always
> basing decisions on real world measurement.
>
> and a simple ssid scan is real world measurement, that helps you understanding what??
>
> btw to remind you:
> this discussion was afair about what would be a wise default,..
>
> which again is 1, 6, or 11
>
> and btw why did you ask this question, if they only answer you want to hear is 7?
>
> Markus
Ug... sorry, this whole discussion has turned a bit silly. My aim was to provide quick snapshots to illustrate the idea I had in mind, to take the discussion beyond a wikipedia page that shows that wifi channels overlap. I would love to see more data to support the 1,6,11 only rule. I'm not convinced that is applicable in all situations. Many yes, all no. If you find this discussion pointless, feel free to ignore me.
Setting the channel to 7 is a minor experiment related to this. I am doing this in the Debian package because AFAIK, there is no current default working setup in the Debian package, and adding some optional settings won't mess with existing setups. It is quite unlikely that the network specified in the Debian package will ever get very widespread. But I do think that it will get more people interested and involved in mesh if it is really easy for them to get started.
.hc
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.olsr.org/pipermail/olsr-dev/attachments/20120629/a4e1421a/attachment.html>
More information about the Olsr-dev
mailing list