[OLPC Networking] Recommendation I3

[Ricardo Carrano]

Hi Oliver,

On Tue, Aug 12, 2008 at 2:24 PM, Oliver Mattos <oliver at mattos.co.uk> wrote:
> Suggestion:
>
> Always use the fastest rate (54) for connections between all peers in the
> mesh.  The only condition that allows the rate to be reduced is when a mesh
> node can't contact any other node, in which case it lowers transmit rate.
>

I find it very tempting to do so, but I am concerned with sparse mesh
scenarios where nodes are not close enough to reach each other at
54Mbps. One example would be children in their homes chatting with a
neighbor. So I believe that an adaptive mechanism would more
appropriate. Do you think differently?



> This algorithm might need some corner cases "fixing" to prevent two
> "seperated" mesh networks where peers in each half can't talk to the other
> half.
>
> On Tue, Aug 12, 2008 at 12:18 AM, Ricardo Carrano <carrano at laptop.org>
> wrote:
>>
>> Hello,
>>
>> Those interested, please take a look at:
>> http://wiki.laptop.org/go/Wireless_Rate_Adaptation_Logic
>>
>> I am collecting comments on the subject and should issue a formal
>> recommendation (which algorithm to implement, for instance) after a
>> period of two weeks collecting comments and after discussing that in
>> person with to M.Bletsas at 1CC by the end of August. For now, the
>> recommendation is "change the rate adaptation logic"
>>
>> Looking forward to reading your comments and suggestions.
>>
>> Cheers!
>> Ricardo
>> _______________________________________________
>> Networking mailing list
>> Networking at lists.laptop.org
>> http://lists.laptop.org/listinfo/networking
>
>