4 ms·

> Or ... people can use the DoD/Mod and assorted other space to get their IPv4 allocations. Someone did the math on this: > Now, average daily assignment rate

> Or ... people can use the DoD/Mod and assorted other space to get their IPv4 allocations. Someone did the math on this: > Now, average daily assignment rates have been running at above 10 /8s per year, for 2010, and approached 15 /8s towards the end. This means any reclamation effort has to recover at least 15 /8s per year just to break even on 2010’s growth. That’s 5.9% of the total IPv4 address space, or 6.8% of the assignable address space. Is it feasible to be able to reclaim that much address space? Even if there were low-hanging fruit to cover the first year of new demand, what about there-after? Worse, demand for address space has been growing supra-linearly, particularly in Asia and Latin America. So it seems highly unlikely that any reclamation project can buy anything more than a years worth of time (and reclamation itself takes time). * https://paul.jakma.org/2011/02/03/why-dont-we-just-reclaim-unused-ipv4-addresses/ https://paul.jakma.org/2011/02/03/why-dont-we-just-reclaim-u... There are 'only four billion IPv4 addresses, and there are eight billion people on the planet. There are just as many smartphones (I have two: personal and work): * https://www.weforum.org/stories/2023/04/charted-there-are-more-phones-than-people-in-the-world/ https://www.weforum.org/stories/2023/04/charted-there-are-mo... Even if you (CG-)NAT an IPv4 address for some number of people, you still need to have IPv4 addresses for public services (web, mail, NTP, etc). There is no scenario where 2^32 addresses is enough for humanity's needs: as some point you need to go to a protocol with more that 32 bits of address space.

Bender 6mo ago
There are 'only four billion IPv4 addresses, and there are eight billion people on the planet. There are just as many smartphones (I have two: personal and work) Unless all of these devices are running a dedicated full time server that must be reachable inbound by everyone this is not required. At any given time "all the people" are not online. That is why DHCP (per ISP) takes care of this. Maybe some day all the people may become terminally online but I would not count on it. Yeah some day IPv6 may be required. Maybe in 100 years or so. IPv4 has plenty of unused allocated addresses that can be ripped away from greedy people. There was a time when ARIN would check to see what was in use and would take back anything people were squatting on. I think the reclamation project works if we dont assume everything has to be reachable as a server. I should add that cell phones (where people are terminally online) were already IPv6 a long time ago for the most part so it's really a non issue. The only risk I see is if someone wanted to start a new massive dedicated server and VPS provider. Most of those are dual stack IPv4+IPv6 now and doing that means clawing some IPv4 space away from those I mentioned earlier.
- throw0101d 6mo ago
  > Unless all of these devices are running a dedicated full time server that must be reachable inbound by everyone this is not required. I think this is a lack of imagination. The fact that (CG-)NAT is in the way could be precluding the development of software that could take advantage of incoming/P2P connections. It's a form of (negative/inverse) survivorship bias: kind of like zoning for only single-family homes and yet saying "no one wants mid-rise towers/apartments as evidenced by the fact no one building them". The current rules/structure preclude any other options. When we went from dial-up speeds to DSL/cable to fibre we were able to have all sorts new applications due to higher bandwidth. Are there classes of applications that we don't / can't have because of NAT? We're stuck with things that often need a central server (TURN/ICE/STUN) and I'd like people to have the ability to explore a more distributed/decentralized Internet.
  - Bender 6mo ago
    I think this is a lack of imagination. No imagination required. P2P works fine if at least 20% to 30% have ports open inbound. 70%+ need not have open inbound ports. Where this could theoretically be a problem is if a specific sub-set of CG-NAT users were the only people seeding and downloading something. This non existent problem can be worked around using a VPN mesh. Tinc is an open source VPN that operates in user-space and while not as fast as Wireguard it can do things Wireguard could never dream of such as user space mesh routing, always discovering the shortest path. The advantage of this is keeping ambulance chasing lawyers off the P2P/VPN mesh. The only imagination required is how to keep the network semi-private. In my experience this is running a semi-private invite-only self hosted forum. In reality none of this is required for P2P however.
    - throw0101d 6mo ago
      > No imagination required. P2P works fine if at least 20% to 30% have ports open inbound. 70%+ need not have open inbound ports. Where this could theoretically be a problem is if a specific sub-set of CG-NAT users were the only people seeding and downloading something. "Seeding"? "Downloading"? I think applications besides BitTorrent could be invented and become popular. Even now, existing things like SIP and WebRTC would probably be much less onerous. > This non existent problem can be worked around using a VPN mesh. Tinc is an open source VPN that operates in user-space and while not as fast as Wireguard it can do things Wireguard could never dream of such as user space mesh routing, always discovering the shortest path. So you're introducing another layer of software because the underlying network does not have the functionality available (just like STUN/TURN/ICE had to be invented to deal with NAT). Here's another idea: have IPv6, and if folks want to have end-to-end encrypted communications, start up an IKEv2 process (that opens a hole for its port via UPNP/PCP), and we have IPsec (which is built into most OSes anyway) encrypted communications opportunistically enabled.

Bender 6mo ago

There are 'only four billion IPv4 addresses, and there are eight billion people on the planet. There are just as many smartphones (I have two: personal and work) Unless all of these devices are running a dedicated full time server that must be reachable inbound by everyone this is not required. At any given time "all the people" are not online. That is why DHCP (per ISP) takes care of this. Maybe some day all the people may become terminally online but I would not count on it. Yeah some day IPv6 may be required. Maybe in 100 years or so. IPv4 has plenty of unused allocated addresses that can be ripped away from greedy people. There was a time when ARIN would check to see what was in use and would take back anything people were squatting on. I think the reclamation project works if we dont assume everything has to be reachable as a server. I should add that cell phones (where people are terminally online) were already IPv6 a long time ago for the most part so it's really a non issue. The only risk I see is if someone wanted to start a new massive dedicated server and VPS provider. Most of those are dual stack IPv4+IPv6 now and doing that means clawing some IPv4 space away from those I mentioned earlier.

throw0101d 6mo ago
> Unless all of these devices are running a dedicated full time server that must be reachable inbound by everyone this is not required. I think this is a lack of imagination. The fact that (CG-)NAT is in the way could be precluding the development of software that could take advantage of incoming/P2P connections. It's a form of (negative/inverse) survivorship bias: kind of like zoning for only single-family homes and yet saying "no one wants mid-rise towers/apartments as evidenced by the fact no one building them". The current rules/structure preclude any other options. When we went from dial-up speeds to DSL/cable to fibre we were able to have all sorts new applications due to higher bandwidth. Are there classes of applications that we don't / can't have because of NAT? We're stuck with things that often need a central server (TURN/ICE/STUN) and I'd like people to have the ability to explore a more distributed/decentralized Internet.
- Bender 6mo ago
  I think this is a lack of imagination. No imagination required. P2P works fine if at least 20% to 30% have ports open inbound. 70%+ need not have open inbound ports. Where this could theoretically be a problem is if a specific sub-set of CG-NAT users were the only people seeding and downloading something. This non existent problem can be worked around using a VPN mesh. Tinc is an open source VPN that operates in user-space and while not as fast as Wireguard it can do things Wireguard could never dream of such as user space mesh routing, always discovering the shortest path. The advantage of this is keeping ambulance chasing lawyers off the P2P/VPN mesh. The only imagination required is how to keep the network semi-private. In my experience this is running a semi-private invite-only self hosted forum. In reality none of this is required for P2P however.
  - throw0101d 6mo ago
    > No imagination required. P2P works fine if at least 20% to 30% have ports open inbound. 70%+ need not have open inbound ports. Where this could theoretically be a problem is if a specific sub-set of CG-NAT users were the only people seeding and downloading something. "Seeding"? "Downloading"? I think applications besides BitTorrent could be invented and become popular. Even now, existing things like SIP and WebRTC would probably be much less onerous. > This non existent problem can be worked around using a VPN mesh. Tinc is an open source VPN that operates in user-space and while not as fast as Wireguard it can do things Wireguard could never dream of such as user space mesh routing, always discovering the shortest path. So you're introducing another layer of software because the underlying network does not have the functionality available (just like STUN/TURN/ICE had to be invented to deal with NAT). Here's another idea: have IPv6, and if folks want to have end-to-end encrypted communications, start up an IKEv2 process (that opens a hole for its port via UPNP/PCP), and we have IPsec (which is built into most OSes anyway) encrypted communications opportunistically enabled.