[MEXT] review of draft-laganier-mext-cga-01

[jouni korhonen <[e-mail]>]

Hello,

In Beijing I volunteered to review draft-laganier-mext-cga-01.

Nits:

o Some RFC2119 inconsistencies
  - s/MUST not/MUST NOT 

o Some typos
  - s/mecanism/mechanism   (several times)
  - s/AGent/Agent
  - s/Crypgraphically/Cryptographically

o Abstract and Introduction are the ~same.. I would expect
  the abstract to be a short abstract.

Comments:

I would summarize the draft as "the ingredients are there
but the presentation is incomplete". Basically, the idea
is imho workable and actually rather nice. I did not spot
any immediate technical show stopper.

However, the draft definitely *needs* an update to make
its idea and protocol aspect to become readable. Current
referencing around few RFCs and assuming the reader to
remember every detail of reused options is quite
overwhelming. At minimum exact pointers to specific sections
in referenced RFCs would help a lot. Especially I would like
to see more text regarding the constriction of the BA message.
For example exactly stating which mobility options go where..
Also the draft should be clear how the proposed solution affects
route optimization. Now it is not clear whether I would
need to do route optimization based on RFC3775 or RFC4866.

Section 8 regarding IPv4 support does not really seem to
fit in this document. Not at least how it is currently
presented.

- Jouni