/> From my point of view, any interconnection technology that requires /> going through a third point (e.g. external OpenVPN server) likely won't There is no reason why you cannot have some VPN servers in datacenters in Australia, and connect them to the system and interconnect them in a mesh fashion. Then everyone can connect to their closest server and you have fast local connectivity, combined with ease of access. When, in addition, you arrange for the subnets involved to be BGP announced on internet, you will have fast internet connectivity as well, without a roundtrip via UCSD. While this is already possible with the current system, it is much more complicated to implement it and it can only be done on a select number of router types. As several people have written, may users are not network architects and have limited knowledge about networking. And there may be many more who are interested in joining the network but have been unable to do so, because of lack of knowledge and/or lack of a suitable internet connection. Making the system easier to access may get us many more participants and after that we may ask ourselves "why did we live with that complicated IPIP system for so long"? Here in the Netherlands, there are 16 registered IPIP gateways (some of them are not actually operational), and 36 active VPN connections with BGP. Plus at this time there are 14 active OpenVPN connections (endpoints not using BGP) out of 220 registered accounts for that. So 50 "new technology" connections vs at most 16 old IPIP connections. That should be clear. There used to be more active connections, it is on a decrease again. Probably after some time the users start asking "what is the benefit of being on this network, what service is it offering that I don't have on plain internet". That is something we should be working on as well. Rob