[tcmtf] BoF feedback

[Luigi Iannone <[e-mail]>]

Hi,

I was thinking a little bit about the turn the BoF took yesterday, going a bit off topic IMHO.

At a certain point people were going to the mic to explain yet another scenario where TCMTF would not work or is not useful.

It is obvious to me that such scenarios exist, but  the goal of TCMTF is not to provide something that will be applied to all packets under a certain size.  

It is also obvious to me that scenarios where TCMTF is beneficial exist as well. Goal of TCMTF should be to provide the right mechanism for those scenarios.

A key point to discuss in TCMTF is how we identify and select the flow we want to be multiplexed on a single compressed tunnel.

All of this to say that we should not focus or scenarios where TCMTF doesn't help but rather to answer the following question:

Are there scenarios and use cases where TCMTF provides benefits?

The answer IMHO is YES so some work should be done.

Just let me add that,  before thinking how to encapsulate we should answer: How do we select the flows we want to encamp? 

So that we are sure that we do not touch traffic that will suffer because of TCMTF.

Just my thoughts after the BoF.

ciao

Luigi