[olsr-dev] Dynamic tunnels to GWs -- increasing MTU [QUESTIONS]

[Aaron Kaplan]

On Apr 19, 2007, at 6:49 PM, Gianni Costanzi wrote:

>> on a radio link with many senders at once you usually don't _want_ to
>> have high MTUs
>> Basically you want to occupy the "time slots" as little as possible.
>
> You're right, but I was thinking about a transparent IP-in-IP
> encapsulation (transparent to OLSR-unaware clients), which would
> require increasing the MTU in order to fit the additional IP header in
> case clients send maximum-sized frames with 1500bytes of payload
> (given that we DO NOT WANT to do IP FRAGMENTATION and we do not want
> to force a change of MTU on the clients).
>
Hm, interesting discussion actually :)


I wonder if that would make any difference?

Because if you have packetloss on the outer tunneling layer, then you  
for sure
will also need retransmits on the inner layer, right?
Regardless if you fragment or not.

;)

So, to sum up my personal answer to your 2 questions:

ad (1): not that I know of
ad (2): yes, the layer 1 and 2 don't like this idea.

ciao,
a.

---
there's no place like 127.0.0.1