[pcp] draft-ietf-pcp-dhcp-05.txt

[Stuart Cheshire <[e-mail]>]

Some comments on draft-ietf-pcp-dhcp-05.txt

I found the document confusing and contradictory. In "Terminology" it says:

>    name may be structured as DNS qualified name or
>    be composed of strings such as can be passed to getaddrinfo
>    (Section 6.1 of [RFC3493]), including address literals, etc.

but "Introduction" says:

>    Motivations for defining PCP option as a name and not an IP address
>    are discussed in Appendix A.

I'm not sure which statement applies.

>    The domain name(s) is encoded as string.

This is ambiguous. What kind of string? It should state explicitly that it's a <character-string> as defined in RFC 1035.

And what can it contain? An FQDN? Does it need to end with a trailing dot? Or does it need to *not* end with a trailing dot, because that would be redundant (but the client MUST add one before passing it to gethostbyname)? Can it contain a dotted-decimal IPv4 literal? An IPv6 literal? In square brackets? Are leading/trailing/interior spaces allowed? Or are they illegal?

>    When several names are included, a space character is used as separator.

If there's more than one, they should be concatenated together as with an RFC 1035 TXT record.

On the other hand, if only FQDNs are allowed, not address literals, then instead of <character-string>s, it would be better to use the already-defined name list format from RFC 3397.

>    Once each Name conveyed in the OPTION_PCP_SERVER option is validated...

How is a name "validated"?

Stuart Cheshire