Dear Christer, Thank you for the review.
Please see inline. Cheers, Med De : Christer Holmberg [mailto:[email protected]] Envoyé : samedi 22 mars 2014 09:49 À : [email protected] Cc : [email protected] Objet : Gen-ART review of draft-ietf-pcp-dhcp-09 I am the assigned Gen-ART reviewer for this draft. For background on Gen-ART, please see the FAQ at < http://wiki.tools.ietf.org/area/gen/trac/wiki/GenArtfaq>. Please wait for direction from your document shepherd or AD before posting a new version of the draft. Document: draft-ietf-pcp-dhcp-09.txt Reviewer: Christer Holmberg Review Date: 22 March 2014 IETF LC End Date: 1 April 2014 IESG Telechat date: N/A Summary: The document is almost ready for publication. I do have a minor issue, and some editorial change suggestions, that the authors may want to address. Major issues: - Minor issues: Q1_A: The Abstract (and, later, also the Introduction) says: "The use of DHCPv4 or DHCPv6 depends on the PCP deployment scenario." I think this is a little unclear. Would it be possible to add some extra text, describing in what type of scenarios the mechanism is applicable? [Med] The decision to use DHCPv4 or DHCPv6 is deployment-specific. The text calls out one particular deployment case where dhcpv6 is used to configure an IPv4 PCP server (see section 5). It is out of scope of the document to list in which cases dhcpv4, dhcpv6, or both will be used to configure pcp servers. I suggest to maintain that sentence as it is. Nits/editorial comments: Q2_1: The Introduction says: "This specification assumes a PCP server is reachable with one or multiple IP addresses. As such, a list of IP addresses can be returned in the PCP server DHCP option." Is that text needed? The document describes how DHCP can be used to return PCP server IP addresses. If there are no PCP servers, no IP addresses should be returned :) [Med] The point of that sentence is to explain why the option is designed to return a "list" and not a single IP address. Q3_1: The Introduction says: "This specification allows returning one or multiple instances of the PCP server DHCP option. This is used as a hint to guide the PCP client when determining whether to send PCP requests to one or multiple PCP servers. For guidelines on how a PCP client can use multiple IP addresses and multiple PCP servers, see [I-D.ietf-pcp-server-selection]." Is the 2nd sentence needed? Why not say: "This specification allows returning one or multiple instances of the PCP server DHCP option. For guidelines on how a PCP client can use multiple IP addresses and multiple PCP servers, see [I-D.ietf-pcp-server-selection]." [Med] The second sentence is useful because it indicates to the PCP client how to determine whether returned IP addresses belong to the same PCP server or distinct one. That sentence will be kept. Q4_2: The Terminology says: "DHCP client (or client)" and "DHCP server (or server)". As the document describes two types of clients (DHCP clients and PCP clients) and two types of servers (DHCP servers and PCP servers) I think the draft text should always explicitly indicate which type of client/server is referred to. I.e. the text should never say "client" or "server", but always indicate whether it's DHCP or PCP. Most of my subsequent comments are related to that. [Med] OK, will update the text accordingly. Q5_3_2: In section 3.2, I suggest to change the section name to "DHCPv6 Client Behavior". [Med] Done. Q6_3_2: In section 3.2, I suggest to say "To discover one or more PCP servers...". [Med] Even if I think the initial wording was fine, I updated the text using your suggested one. Q7_3_2: In section 3.2, I suggest to say "The DHCPv6 client MUST be prepared..." [Med] Done. Q8_3_2: In section 3.2, please add a reference to "IPv4-mapped IPv6 address". [Med] A reference is cited in Section 3.1. It seems you missed that one ;-) Q9_4_2: See my comments, Q5, Q6, Q7 and Q8, but replace "DHCPv6" with "DHCPv4" :) [Med] Same as above. Regards, Christer
_______________________________________________ Gen-art mailing list [email protected] https://www.ietf.org/mailman/listinfo/gen-art
