(Disclaimer: I’m a member of the ARIN AC, but strictly speaking for myself here):
Reading the below, I’m getting the impression that the real need here is some sort of uniform identifier that is globally unique. While there are other forms of identifiers out there in the world (UUIDs, UPC codes, et al), there’s nothing I’m aware of that provides the combination of global uniqueness, hierarchal addressing structure, and number of available addresses. So, it’s not unreasonable to consider using IPv6 address space for this purpose, because it’s already there? Unfortunately, what this means is that we’re attempting to use an addressing format designed for a very specific use case - internet routing - and apply it to problems that may appear on the surface to be similar, aren’t similar at all once you start looking more closely. The most glaring issue I see is that IP addresses were never intended to be *permanent* identifiers for anything - one of the hallmarks of internet routing is that any device can, in theory, be assigned any legal IP address appropriate for it its location within a topology, and send/receive traffic with that address. The mechanism that might be used to assign an IP address to a not-network device is undefined, but if one were to, say, print the address onto a barcode label, then you’ve just *permanently* assigned an address to that object, for all practical purposes. How would you propose that organizations manage assignments that would avoid this? I’m sure there are other issues, some of which have already surfaces, but the above is the most glaring one I see. Effectively, adopting this type of proposal would have the affect of changing the fundamental assumptions on how IP addresses are assigned and routed, and would be a very difficult sell absent a change in Global RIR (or IANA) policy, updates to relevant RFCs, et al. What this proposal *does* raise is the need for a similar mechanism for uniquely identifying products, manufacturing components, et al that provides the similar qualities that an IPv6 address does. I just don’t see any good reasons to mix the two, and plenty of downsides. As such, I cannot support this proposal. Thanks, -Chris Woodfield > On Aug 7, 2024, at 07:02, Wesley <[email protected]> wrote: > > Hi Jonathan > > > This is Wei WANG, the co-proposer of prop-161-v001. > > Thank you for sharing your expertise. I'd like to clarify the rational of the > proposal: we aim to innovate within the current policy framework by expanding > the use of IPv6 addresses, without altering the underlying IPv6 technology > stack. > > > Allocating IPv6 addresses to non-electronic items, is a straightforward > simplied expression of binding a unique IPv6 address to each data object of > non-electronic item. > > > It is reasonable to have a specific domain name printed onto a trade mark, to > assist consumers in obtaining the relevant product information. This can be > regarded as assigning a domain name to a non-electronic item. > > Similarly, our proposal is to directly use the IPv6 address behind the domain > name as the primary ID to routing the user query to the exclusive data object > page of the corresponding item. > > > However, Using IPv6 addresses as the item identifers doesn't mean replacing > other identification schemes. Actually, in practice, IPv6 addresses could be > generated by hashing the upper layer semantic ID to the interface ID/postfix > 64 bits. > > > Introducing IPv6 addresses as the Item ID sets up an effective technical > barrier to the product counterfeiters: > > 1. The ID owner is also the IP owner, ensuring that query traffic is > directed to the correct destination through BGP broadcasting. > > 2. Authenticity is ensured by existing security measures such as RPKI > (Resource Public Key Infrastructure) and CGA (Cryptographically Generated > Addresses). > > 3. Traceability is enhanced by the end-to-end transparency of the > network, providing clear location information for both the source and > destination. > > 4. Flexibility is achieved as IP address administrators can use > techniques like Layer 3 NAT, traffic scheduling, or Layer 7 switching to > direct access requests to any arbitrary IT system, ensuring seamless > collaboration with upper layer identification schemes . > > > In summary, while it is theoretically possible to assign an IPv6 address to > every grain of sand in the world, in practice, this is unnecessary and > impractical for natural sand found in deserts or on beaches. However, once > sand is packaged or transformed into a commercial product, it may require an > IPv6 address for identification and to provide access to the item's > corresponding data object. > > > Best, > > Wei WANG > > _______________________________________________ > SIG-policy - https://mailman.apnic.net/[email protected]/ > To unsubscribe send an email to [email protected]
_______________________________________________ SIG-policy - https://mailman.apnic.net/[email protected]/ To unsubscribe send an email to [email protected]
