Dear Juan Carlos, Carlos, Telemaco, All,

I went over draft-zuniga-dmm-gap-analysis-01 and, although I understand that 
this draft is a reasonable first shot, at times it reads as if it was written 
very hastily.  Overall, much work lies ahead. Please find below my comments and 
suggestions after a cursory review. 


Section 1: Intro
----------------
There is no need to repeat what's in the reqs ID nearly in verbatim. Besides 
the fact that most interested readers would be familiar with this already, you 
may end up with synchronization and consistency issues, which need to be taken 
care of on a continuous basis till -reqs becomes an RFC. Even so, I propose to 
rewrite the intro so that it matches the scope and target of this document, and 
remove the repetition of the reqs text.


Section 2: "Practices"
----------------------
The term "DMM environment" is not actually defined; ditto the term "DMM mode of 
operation".

If you refer to dynamically, on-demand change of the anchor point, then it's 
better to say so explicitly. If, on the other hand, this latter term refers to 
core offload (as in "thus providing some form of distribution, since the 
traffic is locally route[d]"), alternative paths, etc. then it's better to say 
so as well, in the respective paragraphs.

With respect to Figures 1-4, I like that you tried to visually display various 
solutions. Unfortunately, they mostly fail to convey the inherent differences. 
Moreover, the text does not take advantage of these figures whatsoever after 
they are introduced. Further, I am not sure what exactly does the cloud-like 
container for the (different) HA (flavors) and LMA is supposed to tell me. In 
short, I would consider improving the figures significantly as they can be 
powerful in clearly showing what are the "gaps" we're after. But left as-is, 
they provide little value; we're better off, for instance, with a ref to RFC 
6301.

Fig. 2 could be merged with Fig. 1, as it adds little further info.

Somewhere midway in section 2.2.x, the text gets a bit rough. I hope the next 
version will take care of this. Then, somehow the text starts talking about 
possible solutions while still in the "Practices" part (for example the 
recurring references to an inter-LMA mobility protocol).

I'm also a bit puzzled about the sudden appearance of 3GPP stuff in section 
2.2.2 onwards. Forget about how it actually comes together as smooth text (it 
does not). I'm more interested in how it provides a coherent line of 
argumentation in each 2.x subsection and how this relates to the gap analysis 
in section 3. Also, there is a bit of mix-and-match going on, e.g. section 
2.2.5: how's multihoming related with the previous stuff?


Section 3: "Gap analysis"
-------------------------
>From section 3.1.2 onwards, the text is, at times, just rough/copy-paste 
>notes. This is particularly the case for 3.1.3-6 and 3.2.3-6. The format 
>adopted for section 3 is also more like a sketch of a gap analysis.


Overall comments
----------------
The document is over-structured and does not lead to new insights esp. in 
section 3.

A conclusions section that pulls everything together is clearly missing.

Finally, in the references section, I'm not sure why certain RFCs are 
considered "normative" for this draft, esp. since no requirements language is 
specified (there's no ref to RFC 2119 for that matter). A ref that perhaps 
should be included is RFC 6301. Ditto, refs for DIDA, OPIIS, and MAPCON are 
missing. Are all listed refs actually used?


A couple of nits and editorial suggestions follow:

---
s/the MN two different temporal/the MN has two different temporary

s/used the RCoA/uses the RCoA

s/locally router at/locally routed at

s/a point of view of routing/the routing point of view

s/considered as a mean to/considered as means to

s/provides a better/provides for better

s/anchored and a/anchored at a

s/[REF to flow mob draft]/ <actual reference>

s/signaling overhead/signaling overhead

s/only performed/only be performed

s/This approach include/This approach includes/g

Define acronyms on first appearance. Not all readers are familiar with them.
----

Best Regards,

Kostas



_______________________________________________
dmm mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmm

Reply via email to