Re: [v6ops] Comments on 464xlat-00

[Washam Fan <[e-mail]>]

Hi,

>> 1. section 7.4 mentioned DNS proxy for both IPv4 and IPv6 hosts, i
>> don't know why DNS proxy for IPv6 hosts is needed.
>
>
> The 464XLAT does not require DNS Proxy for both IPv4 and IPv6 hosts.
> IPv6 host may directly query an IPv6 DNS server.  IPv4 host may also
> directly query an IPv4 DNS server through 464XLAT translation.
> DNS Proxy is useful function for IPv4 host so the host can not easily
> discover an IPv4 DNS server address.

Got it. But it seemed to me that such DNS Proxy is required when I
read section 7.4.

>> 2. I think the author should elaborate on what the configuration
>> function is like in section 7.6
>
>
> I do not know where your consideraton about this configuration
> function is.
> Could you please specifically indicate about your consideraton?
> I think we can make a appropriate document by yours.
>
i was referring to the configuration function which allows the PLAT
and CLAT not to include the Fragment Header.

>> 3. i don't know what is the meaning for the text under section 7.5. I
>> think NDP should always perform while SLAAC perform only if the prefix
>> equals to 64.
>
>
> If CLAT gets a single /64 prefix from WAN interface, it has to perform
> ND for 464XLAT IPv6 addresses(It means that IPv4 host under CLAT),
> and it has to perform ND-Proxy for IPv6 host under CLAT.
>

Do you mean ND for the translated ipv6 addresses where ipv4 addresses
assigned to the hosts under CLAT are embeded? If yes, then i think you
need to clarify in this section because this section is too vague.

My point was ND can perform even the prefix != 64, cant it? And when
you say ND, are you referring to DAD or full functionality of ND?

>> 4. I don't know if the draft should cover the host1-CLAT1-CLAT2-host2
>> communication case.
>
>
> This 464XLAT architecture is hub and spoke model.
>
Would you mind clarify this scope in somewhere in the draft?

Thanks,
washam