OwlNet

Ok, that’s it, I’m tired of the Internet. The ipv4 crash? Speculative hoarding of netblocks, no doubt. Net neutrality? Uhh, seriously? Egypt’s government running scared and cutting off electronic communications to delay the inevitable? Verisign and the U.S. government playing around with wielding imaginary and annoying power with the DNS system? Yeah… No more.

For years, there have been mesh networks run in various places, usually a small local system. These networks can use a variety of link types; wired Ethernet, wireless radio, point to point microwave, fiber, etc. These networks are self-healing in the event of link failures or obstructions.

This is in sharp contrast to the topologies in most current commercial networks used to connect end users to the Internet, in which cost engineering has dictated that a hub and spoke model be used.

This usually kind of works, but has some major flaws. First, traffic may be needlessly routed over very long paths between major backbone links simply because two geographically close locations do not have any option for a direct link.

Second, if major backbone providers experience link failures or simply decide to “de-peer”, massive routing problems can affect millions of users. I still remember the day Level3 and Cogent came apart… I couldn’t get my e-mail for the better part of a week, as the routes were simply not there.

So, enter Owlnet. The name may or may not have been inspired by the works of a certain author whose series of novels stimulated the imaginations of a world alike, and brought her from rags to riches. By the way, if she says he is, Dumbledore is gay. Arguing this is hilariously illogical.

Owlnet shall consist of a networking layer, using a mesh network that may be deployed using whatever network hardware is available, and a services infrastructure. Wireless lan, ethernet, and even tunnels over the existing Internet may be used. (Ideally, the tunnels will be phased out as the system gains coverage.)

Dare I say that, at least for my point of view, is the easy part. The goal is to have firmware and software available to allow Owlnet nodes to be run on flashable home routers in the sub-$100 proce range, as well as on PC and server architectures.

The services infrastructure will require quite the creative planning. One of the goals I see having is an independent dns system, as well as the ability to route out of the system via exit nodes to the existing Internet, similar to Tor’s onion router system. This could be set to different personalities, everywhere from “use the closest and most convenient node” to “INVISIBILITY CLOAK!”.

The independent dns would be an interesting challenge. Ideally it should never fall under the same curse of greed that has afftected the system used on the Internet, and should be reliable and easy to use for registering resources, yet resistant to abuse. For instance, it should prevent a user from easily holding large numbers of domains for profit from resale, or squatting on existing names and trademarks. Unfortunately, I also want to avoid putting one small group in control, so this does create a bit of a paradox when it comes to how the system should be managed.

IPv6 should be used on this system. We all know by now what happens to a small namespace.

Discovery and search are two other interesting points to explore. I want the system to have a phonebook of sorts for personal contact, as I envision connecting ip telephony and instant messaging as two major applications. It should be possible to find someone via personal connections through a unique identifier, but also by more traditional means like by name and location, or via personal networks as is done with Facebook and other social networks.

The ultimate goal is to have a network that’s very usable and not dependent on any outside infrastructure. The nodes could literally be little solar/wind powered units you can toss in high locations and let it all run. Where a gateway to the existing Internet is desired, they could be connected via home or office connections as needed.

Come to think of it, this sounds a lot like HackMiami’s “Post-Apocalyptic Communications” concept. Well… let’s get it going 🙂