I want a network that scales to trillions (or more) of top-level nodes (rather than the ~2 million supported by IP and BGP.) I want everyone to be able to host a router.
I'd go source-routed isochronous streams, rather than address-routed asynchronous packets.
I haven't updated my blog in a few years, but I'm still working on building the above when I have the time. (IsoGrid.org)
I'm not much of a football fan, but I would really love it if the Seahawks were owned by a charity. It would make rooting for the home team much more meaningful!
You say it's easy to imagine bad scenarios. Then state them, with sources.
I've read the exact opposite: That it's hard to find any _real_ risks for eliminating human-biting mosquitoes. I define real as killing near 3k children a day.
Hi, I'm the main developer behind the IsoGrid Protocol. Let me know if you have any questions!
I'm looking for folks who might be interested in helping out. I was raised under Windows, but I want to use free (as in freedom) software, so integration has been a bit slow going as I learn how to use Linux.
I agree with you, I would also rather be reading about a stable and popular implementation of a scalable mesh network. However, there is no such thing. TCP/IP is the closest I've seen, but I think we can do better.
Nothing has changed technically. TCP/IP has consistently caused the centralization of power. The 90's Geeks didn't build the protocol, DARPA did (decades prior).
We wanted anyone to be be able to be a first-class citizen of the Net. But we ended up beholden to our Monopoly ISPs. TCP/IP only gives first-class status to ISPs, not to us. BGP doesn't scale if we all tried to have our own first-class routers.
Yes, things are getting better in all sorts of industries; a similar boom occurred with the wide-scale deployment of the 19th century Railroads. But things got better still with the deployment of automobile roads.
I propose we do the same with a globally scalable mesh network.
Right. Centralization is a physical fact of the Internet as we now know it.
AFAIK though, DARPA wanted a design that could survive nuclear war due to its' distributed nature. Has the Internet and in particular TCP/IP, been successful in that regard? It would appear not.
TCP/IP centralizes. Why am I stuck with one ISP? Scalable Mesh is the solution. I just GPL'd my project I'm working to build: https://GitHub.com/IsoGrid/IsoSwitch
No. TCP/IP can work just fine over multiple ISPs, Google "asymmetric routing". Also, look into cloud load balancers like cloudflare, then realize you can use A records to load balance over more than one of those with multiple different backends. The problem is physical diversity of connection at any one location, not some perceived weakness in TCP/IP.
I assert that TCP/IP is actually the cause of "[lack of] physical diversity of connection in any one location".
A second (or third) ISP isn't going to suddenly show up at my house because it doesn't make economic sense. In the same way that two railroad lines don't get built between the same two cities.
A real scalable mesh protocol could change the economic calculation.
That are two things that really killed the peer-to-peer nature of the Internet---NAT and firewalls. Both work to make a true peer-to-peer connection difficult.
My wife and I have been working on this for a few years. It's still in the early stages (we haven't built any HW), but we at least have a HW architecture plan.
I'd go source-routed isochronous streams, rather than address-routed asynchronous packets.
I haven't updated my blog in a few years, but I'm still working on building the above when I have the time. (IsoGrid.org)