Dear WG,
https://www.ripe.net/participate/ripe/wg/ipv6/documents/itu-ipv6refmodel
I would like to thank SG20 for the liaison statement and the opportunity to comment on Draft Recommendation Y.IPv6RefModel "Reference model of IPv6 subnet addressing plan for Internet of things deployment" from Q3/20 e-meeting December 2017. The document suggests an "optional reference model of an IPv6 addressing plan for Internet of things (IoT) deployment by smart cities, public administrations and companies". However, I am missing any identification of the special needs of these target audiences, any explanation of specific needs of "IoT" (with that being resollved per ITU-T Y.2060 as a "global infrastructure"), or the benefits of a unified addressing plan at all. The draft document lacks precision in terminology (e.g., it talks about "routing prefixes" rather than "address space"). The document refers to IETF "standards" dealing with "IoT", but fails to give an explanation for a need for an addressing plan arising from the use or deployment of those standards/protocols. In addition, there is no reason given why SG20 would be the appropriate venue for this activity, rather than the IETF or the RIRs and their attached operator communities. Section 7 "Preventing a New Digital Divide" suggests that the lack of a reference model for an addressing plan was detrimental to IPv6 deployment, probably based on region - without giving supporting facts. Also, the plans proposed later would not be sufficient to justify address assignments (or allocations) beyond those sizes easily available today, nor is there any support for an argument that "shortage of v6 addresses" would be an obstacle for deployment. The plans also do not help in utilizing the address space and deploying real v6 networks (see below). Section 8 adds "smart buildings" to the target audience without further explanation. The section then lists a number of IPv6 deployments, criteria for inclusion in the list being unclear. There is also no link the the draft document, neither by those examples having informed the proposed plans, nor by the the proposed plans having been instrumental in the operational reality of any of the examples. Section 9 does not correctly represent the role of RIRs and LIRs (or NIRs, where applicable). The plans in section 11, finally, deal far more with the "non IoT" parts of the network (DMZ, servers, regular LAN) without justifying the recommendations based on operational experience. Also, while there is a postulated size for the address space for "IoT", the size and need is not justified, nor is there any inner differentiation for systems or objects in the "IoT" category. Assigning addresses per building or floor is already outdated in today's networks. In my view, the document clearly fails to motivate the need for a reference model for an addressing plan and any special role for "IoT" in any such plan (read: there is no problem statement). Therefore, it is impossible to evaluate the proposed solutions. That said, the proposed plans are heavily encumbered by an unmotivated "consistency" with IPv4 addressing. -Peter -- Peter Koch | | pk@DENIC.DE DENIC eG | | +49 69 27235-0 Kaiserstraße 75-77 | | 60329 Frankfurt am Main | | https://www.DENIC.DE ------------------------------------------------------------------------- Eingetr. Nr. 770 im Genossenschaftsregister Amtsgericht Frankfurt am Main Vorstand: Helga Krüger, Martin Küchenthal, Andreas Musielak, Dr. Jörg Schweiger Vorsitzender des Aufsichtsrats: Thomas Keller