[nwcrg] Review draft-yang-nwcrg-bats-03

[Emmanuel Lochin <[e-mail]>]

Dear all,

After carefully reading "BATS Coding Scheme for Multi-hop Data 
Transport", I have no specific comment on it.
The draft is well explained and self-contained. It appears almost clear 
to me concerning the implementation of BATS.

Just some slight comments/questions :

- Being a little bit fussy, I would just rephrase "Existing network 
protocols like TCP/IP use end-to-end retransmission and 
store-and-forward at the relays" by "Existing transport protocols like 
TCP use end-to-end retransmission while network protocols such as IP 
might enable store-and-forward at the relays". I understand that you 
want to point out that TCP/IP has the same objective that the 
combination of both inner/outer BATS codes while less efficient over 
multihop;

- Correct "Suppose that a pseudorandom number generator Rand() which 
generate an unsigned integer of 32 bit" -> "generates";

I understand that you need one BATS session per flow at the outer code 
and you write that "The inner code comprises (random) linear network 
coding applied on the coded packets belonging to the same batch.". So my 
questions are:

- Does a batch similar to an encoding window as defined in RFC8406?

- Does the inner code must differentiate each microflow before 
re-encoding? If yes, without re-encoding (pure e2e exchange) does the 
Batch_ID field still mandatory?

- Is there a relationship between the e2e microflow and the Batch_ID? 
Meaning a Batch_ID always identifies a source/sender or a one of its 
microflow?

Regards,

Emmanuel


-- 
Emmanuel LOCHIN
ENAC - Ecole Nationale de l'Aviation Civile
7, avenue Edouard Belin CS 54005, 31055 Toulouse Cedex 4 France
http://www.enac.fr
https://elochin.github.io/