Re: [DNA] Comment on draft-jinchoi-dna-goals-00

[JinHyeock Choi <athene@sait.samsung.co.kr>]

Dear Daniel.

Thanks for useful comments. Kindly find in line comments. 
 
> Firstly, I quite agree it is a valuable draft at this stage 
> and here is my comments as well as minor editorials; 
> 
> [1]
> This draft uses an ambiguous terminology as "node". 
> As indicated in RFC, node means both host and router
> however AFAIC almost node wrote in this draft just mean 
> host not router. So you have to clarify that altogether. 

I see. I will make modification. 

> "L3 link" is not proper since link as also indicated in RFC
> means the layer immediately below IPv6. IMO L3 or IP layer
> is correct.

In this draft, the term 'L3 link' means the same as the term 'link' does in 
RFC 2461. The reason we add 'L3' is to differentiate 'L2 link' and 'L3 
link'. An L3 link may consist of multiple L2 links. 

Here is an example. It's common for several APs to share one access 
router. Each AP forms different BSS, hence different L2 link. When a 
mobile nodes moves from one AP to another, it changes L2 link but
still remains at the same L3 link/ subnet.  

We were not entirely happy about the term 'L3 link' and, in mailing list, 
contemplated the alternatives. But we couldn't find an appropriate one 
in time. We may change the term later, if suitable one appears. 

> [3] 
> >1) Link local scope of Router Address  
> >So if it happens that two different router interfaces have the 
> >same link-local address, a host can't detect that it has 
> >moved from one interface to another by checking the router 
> >address in Router Advertisement messages. 
> 
> I don't think above case is a real model, why two different router
> interfaces are used with the same MAC ? did you have any 
> experience on that ?

Link-local not link-layer. 
 
> [4] 
> >Alternatively a host can monitor periodic RA beacons.
> 
> RA beacons should be replaced with RAs. Beacon seems 
> like a low layer terminology not IP layer.

I see. 
 
> [5]
> >To design adequate DNA schemes, we'd better investigate how
> >to 1) design a RA optimized for DNA and 
> >2) deliver a DNA optimized RA to a host sufficiently fast.  
> 
> It is so ambiguous to make sense clearly. I think (1) is focused
> on both host and router and (2) is focused on router only. correct ?

No. Both router and host need to be involved in them. At least host 
needs to understand what that RA means. 

> [6]
> >G3. In the case where a host has not changed link or subnet, 
> >IP configuration change should not occur. 
> 
> Don't you consider renumbering case without link change ? in 
> this case, host wil change its IP configuration regarding
> new prefixes and addresses.

I didn't consider renumbering case. It seemed to me out of DNA 
scope to deal with renumbering related problems.  
 
> ========================
> minor editorials;
> 
> [1]
> A host can check link change to determine the validity of 
> it IP configuration. 
> 
> s/it IP configuration/its IP configuration
> 
> [2]
> The validity of other IP subnet related configuration is 
> implied by Router Advertisements, if RA reception implies 
> subnet change. Particularly, Dynamic Host Configuration 
> protocols [DHCPv6] are explicitly indicated in RA messages, 
> 
> s/protocols/protocol or protocol for IPv6
> s/are/is
> 
> [3]
> all around
> s/new (IP) configuration/a new (IP) configuration 
 
Thanks for useful comments which made the draft better.

Best Regards

JinHyeock