> If we don't want to share a common transmission resource, then why do we need > globally unique addresses to use in IP packet headers? Locally unique > addresses would do just as well.
Just to answer this question specifically. We may not need globally unique addresses. But I need a unique address for anyone I want to talk to and I don't care what transmission networks my packets traverse. Therefore, we need unique addresses. However, lets say an address is 24 bits long and we use a random number to generate the address. It is unlikely that there will be an address collision for all the things I want to talk to. So to me I get my unique address. Is it globally unique, well no, but maybe it doesn't have to be. But there will be hosts that want to talk to everyone in the world or at least beyond an address collision domain, so we default for the desire to want/need globally unique addresses. So simply using a random number generator for an IPv6 address may get us there and work sufficiently. Comments? > This question could be posed in the context of the evolution of NAT > deployments in today's Internet. NATs were originally seen as a way for edge > networks to share a single provider IP address across multiple devices own > the home Or said another way, NATs gave us 48-bit addresses that were guaranteed globally unique. > network. This is still the case, but address scarcity has also pushed the > access ISP to deploy NATs at the external edge of the access network, using > private addresses comprehensively within the internal network infrastructure. > This Also Geoff, people talk and desire address privacy. Well when there are 3 NATs between you and me, my source address for packets addressed to you are pretty well obfuscated (the second NAT just destroyed the identity of my access network provider, and the third NAT destroyed the identity the transit network provider). So we got this feature by stumbling into it! > provides greater address utilisation efficiencies, allowing the access > network to stretch the public IPv4 addresses across a greater number of end > clients. But if the bulk of all data delivered to customers is now sourced > from a local data centre that houses the local points of presence from the > content distribution networks, then what would happen to Yes, but the TCP acks going back to them are obfuscated/NAT-traversed source addresses. > the pressure on the access network's IP address pool if the NAT was pushed > inside the local data centre? Or to phrase it in the other direction, what > would happen if every content network had a point presence on the "inside" of > each access ISP's network? From the perspective of the content provider > nothing changes. The client IP address is relative to the local point of > presence, so the same local IP addresses can be used in multiple points of > presence with no impact on this model. But from the perspective of the demand > for globally unique IP addresses a lot has just happened. There is no > residual need for them! But the high-order 32-bits of that 48-bit address must be unique. And we know the allocation of those bits comes from outside of that host sending packets (non local matter). Where the low-order 16-bits can be a local matter and hence do not have to concern itself when TCP/UDP ports number could collide among 2 hosts talking to each other. > This line of thought takes the edge innovation observation and just pushes it > a but further! > > regards, Joel interpreted this as bad news and was hoping Geoff would be wrong. But what if Google wanted packets to come into my host with an obfuscated source address. They might want that. The solution is ugly and maybe costly but its not like we would be trying a brand new technology since NATs have been around forever. To be clear, I am not encouraging this. Dino _______________________________________________ Int-area mailing list Int-area@ietf.org https://www.ietf.org/mailman/listinfo/int-area
