Re: Flow label versus Extension header - protocol itself

[marcelo bagnulo braun <marcelo@it.uc3m.es>]

Hi Margaret,

El 19/05/2005, a las 23:29, Margaret Wasserman escribió:
...
> My concern (once upon a time) with extension headers is that the use 
> of extension headers on some packets and not others effectively 
> presents a variable MTU to upper layers.  We need to make sure that 
> this is properly handled to avoid bad interactions with PMTU discovery 
> code.

Right, but note that the extension header is included after a new path 
is used. I mean, as long as the communication flows through the initial 
paths (i.e. the ULIDs are used as locators) there is no need of the 
extension header.
Once there is a failure, we need to change the path, which means 
changing locators and then we need to include the header in data 
packet.

so considering that the extension header is included as a result of 
changing paths, it may be reasonable to expect a change in the MTU 
anyway (because of the path change)

Regards, marcelo


> Margaret