[olsr-dev] Fwd: Frottle and Meshes

[Jim Thompson]

Tim Schmidt wrote:

>That looks interesting, and possibly more efficient than my scheme
>(although without an accurate simulation of each I can't tell). 
>However, it's also drastically more complex.  My scheme would require
>only the use of the ETX table already maintained by the OLSR ETX
>plugin, a timeout timer, and some code to perform handshaking and
>token passing...  Important features of my scheme are:
>
>- no node is a 'master'
>- no node needs topology knowledge beyond who it can hear / who can
>hear it (already provided by ETX plugin)
>- traffic isn't corralled into 'rings', nodes dynamically determine
>if, when, and who it's OK to talk to based on the simple, predictable
>behavior of all other nodes it can hear.
>  
>
just because you can't hear a node doesn't mean it can't hear you.

Also, I'm curious to know if you've thought about recovery, should the node
thats holding the 'permission slip' become unreachable.

>I'm not saying my scheme is perfect...  Hell, I don't even know if
>it'll work.  Even if it does it's a solution at the wrong layer.  But
>I think it would be easy for a talented person to code up quickly, and
>I'd like to see it tested.
>  
>
so write the perl and I'll see about convering it to 'C'.

One of the *fundamental* problems with current 'mesh networks', however,
is the 802.11 MAC.

Jim

>--tim
>
>On 3/2/06, Jim Thompson <(spam-protected)> wrote:
>  
>
>>You might investigate the "wireless token ring protocol"
>>http://www.eecs.berkeley.edu/~ergen/docs/WTRP_SCI02.pdf
>>http://www.eecs.berkeley.edu/~ergen/docs/main.pdf
>>http://www.eecs.berkeley.edu/~ergen/WTRP/
>>
>>I've considered coding this as a replacement for net80211 (the 802.11
>>MAC portion of the 'madwifi' driver), rather than running over the top
>>of the 802.11 protocol.
>>
>>OLSR could still run on top of WTRP, tying together multiple rings.
>>
>>Jim
>>    
>>
>
>_______________________________________________
>olsr-dev mailing list
>(spam-protected)
>https://www.olsr.org/mailman/listinfo/olsr-dev
>  
>