Hello Ruben, please find my answers below: Yes, we will still use 1%. But if we exclude half of our users, then we’ll use 0.5%, which is much less (about half!). We can’t force people to use the space, we just try to help them use it for what they want to. Hopefully, if we do that, we can grow to 2%! I do not believe that, but that’s another discussion: one I’m not going to have ;) On a serious note, right now, today, Friday, we have more users that are on the Intranet and want to remain there than the rest of the network. Germany has most of the users of 44/8 and they all seem to want to join this Intranet. If anything, the people that want to use BGP are the minority. Yet we still want to serve them, as they should too be able to what they want with the IP addresses. Yes, the IPIP Mesh (and setting up BGP) are not easy, and I definitely believe that we will see more use if we make it simpler. And this is what the PoP will try to address. We’re working on it, just as we work on this policy. In this proposal we reserve enough space for both use cases. If any of them reaches 100% (44.0/10 or 44.128/10), then we will proceed to give out space from 44.64/10. Until one of them is close to that, we do not consider resource exhaustion a problem. This is not technically accurate. A user that wants to be on both networks can add a static route to 44.128/10 and then connect to 44.0/9 through their ISP, over the normal Internet. If they want to use a 44-address to reach BGP hosts then they can add another static route to 44.0/9 to a service that can connect them, such as the ARDC PoPs. Then this service can statelessly and effortlessly netmap their 128 space to 0. Alternatively, this user can make use of ARDC’s (or anyone else’s) VPN profiles to receive a dynamic IP within both networks (by making two VPN connections) and then depending on the destination their computer or phone will automatically choose the correct IP. They don’t even need any allocation, or any set up. Just two VPN profiles on their phone, and they can reach anything and everything without any problems. This works because OpenVPN or Wireguard or […] can only route specific networks over the connection by sending a route to the user. This proposal does not make any changes to the technical means a site is connected with. We leave this up to the user. All these operators of repeaters can still connect to the Internet using NAT, netmap, IPv6, or any other way the operators want to! First of all I think that you are being excessive. We are not destroying anything and we do not want to cause any harm to anyone. We are also hams and networking hams. If you are just building your network, you can decide on which part you prefer to be, and receive an allocation there. You are not affected by any renumbering. As I said in an e-mail to Mario a bit before, we don’t think that the only people that use the network are the ones on the mailing list. We also don’t expect every network user to be on the list. And finally, by definition the people that are against something will speak up much more than the people in favor. So I don’t think what you say about most people disagreeing is true. The current TAC supports this design because we want to set the policy, and not impose any technical changes to anyone. THIS IS NOT A TECHNICAL PROBLEM WE ARE TRYING TO SOLVE! That said, it’s unlikely that this can be solved with a technical solution. It’s a people problem. We need to make room for more people and make it AS EASY AS POSSIBLE for them to do WHAT THEY WANT, and HOWEVER they want to do this, ANYWHERE where they want to do this. Forcing people to do anything of technical nature goes against that. A lot of times technical people think they can solve every problem there is out there with a technical solution. This couldn’t be farther than the truth. We can easily get dragged into designing the perfect network with 200 PoPs and 50 VPN technologies, and we go over the fact that some problems are simply community issues, people issues, etc. I doubt that anyone can come up with a technical PoP infrastructure that addresses every possible need and use case that people might want to do. So instead of trying to have a solution that does it all or that it forces people to follow a certain path, we just opted for a solution that guarantees two things, and from that point on anyone can build anything they want on top of them. Try to imagine this solution in a world where the ARDC PoP system will never arrive. We will never deliver, or we will and then we will turn it down 6 months later. Do not rely on some Holy PoPs, Masters of Everything that are supposed to coordinate everything in a centralized fashion. Try to have a more distributed and mesh system in mind where everyone can connect to everyone else and exchange packets without any central authority to tell you what to do and when to do it and how to do it. We do not know or care about what the Intranet will have. We do not know or care about what the Internet will have. We have no control over any of that, and this is up to the users. The Intranet will have whatever you set up for it to have, and the Internet will have whatever you set up for it have. We just have a lot of people that tell us that they want this, and if they say so, we are trying to give them what they want. We also have a lot of people that tell us they want free IPv4, and since they so, we are trying to give them what they want as well. Just become I belong to the second category doesn’t mean I should prevent the people in the first from having what they want. How would you feel if we had a poll on “which use case do you want for 44/8” and people voted “Intranet”? Would you like it for us to make the entire space Intranet and prevent all BGP advertisements and revoke all LOAs? Would you like 44/8 to disappear from the Internet, and ARDC to tell people who want free IPv4 to go and buy it? If you don’t like this thing, why force other people to do it the same way you are? Because if we have a poll, there is a real chance that people will vote Intranet only. Does that mean that the ARDC should take all your space? We believe that every person should have their space to do whatever they like. Intranet people should have their space, Internet people should have their space, and if a new use case exists in the future, we also want to be able to give these people their space as well. It’s a huge space and we can all fit inside and play nicely without being limited by others. It’s okay! As long as we stay calm and civil here it should all be fine :) We are always interested in hearing more opinions. We now have the problem of “we’ve been discussing this for 5 months, and we have everything thought of and in our head, and we just have to present it to the people”. Of course, there’s always the chance we missed something, and that’s why we want to have this public discussion. It’s just that a lot of points that are brought up were also brought up internally. I myself even asked if using public IP space for an Intranet is a good idea. I also did not like it the first time I heard of it. But I always tried to keep an open mind, look at the problem, and then evaluate this possibility. And the more I thought about it, the more it made sense and seemed as the only way to guarantee users of both networks a global uniqueness. But that wasn’t enough. We then studied how people currently use the space, how much we have, how many new allocation requests we get daily, etc. This gave us another piece of data: we have enough addresses. We then kept looking for other solutions, and by the time we had to decide, we looked at many things, we thought of many ways, and we had to combine it all and arrive at a final choice. This wasn’t easy. We had to make tradeoffs and we had to make some users unhappy. I wish we could avoid it, but any choice we made would lead to some people being unhappy. We could only try to minimize this amount of people, and even then try to give them as much support as we can to ease the pain. We understand that a /10 is very expensive, we’ve looked at the market price, we compared the price per IP depending on size (to see how much more a /10 is worth over 64 * /16 for example). Eventually we decided that we prefer to use more IPv4 addresses instead of making more actual humans unhappy. A lot of people may disagree with this decision, and they would prefer to see more unhappy people than IPv4 being used like that. That’s also fine, and we accept that. Our goal here is not to convince people. We’re just trying to help you understand why we made *this* decision and not the *other*, and why we stand by it. But we always do this with an open mind, and without looking at our personal interest. We are open to the fact that we could have missed something entirely. It may appear as we have already set our mind, but I can assure that we’re willing to reevaluate at any time. It’s just that we’ve been through most of these arguments internally and we thought about it, and we already have an answer that is good enough for us. Finally, the TAC is working on many other things as well, including PoPs, IPv6, RPKI, etc. but this just happened to be the first one that came out of the pipeline. We have done work in parallel, and it will slowly start to get to this mailing list for your feedback. Please also keep in mind that the TAC has no budget or staff assigned to it, which means that we can’t approve expenses, anything we do has to be approved by the Board (and ARDC Staff). We also have no staff allocation to us to work on things. We have an advisory role where we tell ARDC how we think they should build a PoP system (and that they should) and then it’s up to ARDC to build and maintain this. It’s not us that will be writing the code or administering it. We can experiment on our own time and come up with some advice (e.g. PPTP did not work, use L2TP), but eventually it is ARDC and its staff and contractors that will deliver the final product. Antonis