Background and History

Distributed Network 42 is an agglomeration of networking technologies, each built and developed often over multiple decades, and the network itself has already been around for almost a decade (at least), having started to become organized in mid-2012 (Home, n.d.). From what limited historical information there is about the network, I’ve come to learn that the network first started out with just a few members tinkering together. Eventually, they set up a wiki and a couple years later a registry to coordinate resources for new members. It eventually also became a way for hackerspaces to connect to each other, and to make it possible for tinkerers to connect to their local hackerspaces remotely to work on projects (Home - Hackerspaces, n.d.). Over time, other similar networks have popped up, and many have become accessible via DN42, when their goals align (for example, ChaosVPN allows hackers and hackerspaces to connect to each other and is therefore interconnected with DN42). Beyond this, the project was and still is a community-owned effort, and as such its history is made of too many projects to list, and thousands of individuals each doing using the network for their own reasons as part of the larger community. Very little historical information on the network is recorded by anyone, so it’s difficult to pinpoint any turning points in the network’s history. It’s simply organically grown into what it is today.

The technologies in use on DN42 rely on the innovation of technologists from many years ago. One of the main tools which enables such a massive networking project to work is the Border Gateway Protocol (BGP), which was first suggested in June 1989 in Request for Comments (RFC) 1105 (RFC 1105, 1989). The protocol, when released, allowed for a scalable way to discover and choose routes from one location on the developing internet to another. This experimental "version 1" was then obsoleted by a newer standard in June 1990 (RFC 1163, 1990), then again in October 1991 with BGP-3 (RFC 1267, 1991). While BGP-3 is not considered obsolete, the current version in use today is BGP-4, which was standardized in January 2006 with RFC 4271 (RFC 4271, 2006).

Even before 1989, there were already similar protocols in use. One of the first was the Gateway-to-Gateway Protocol (GGP), which naively based its routing decisions only on the amount of hops from one Autonomous System (AS) to another (RFC 823, 1982). The Exterior Gateway Protocol (EGP) superseded this protocol, and was standardized just a couple years after the GGP, in April 1984 (RFC 904, 1984). The protocol that laid much of the groundwork for BGP, called the Routing Information Protocol (RIP), was developed and standardized in RFC 1058 in June 1988 (RFC 1058, 1988). It may have taken over a decade after the ARPANET (the precursor to the internet) was created for GGP to be developed, but since then protocols have come and gone, and BGP remains one of the most reliable and widespread protocols used today (The History of Border Gateway Protocol, 2020). Without it, it is difficult to say whether the internet, much less the DN42 project, would ever have reached the magnitude it has.

The other major innovation which enabled DN42 to thrive was the creation of VPNs and tunnels. Whereas AS’s on the internet interconnect using fiber optic cables and copper wires, laying a fiber optic cable from your home to someone else’s home in another continent would be impossible, or at least make the cost of entry to a hobbyist project like DN42 so high only billionaires could participate. Thankfully, there is a way to securely connect to anyone using existing infrastructure like the internet. The first way to do this was described as part of the Distance Vector Multicast Routing Protocol (RFC 1075, 1988), which implemented "tunneling" so that routers could communicate with each other if multicast networking was not supported. This introduced the idea of tunneling one protocol within another, and was eventually standardized as IP in IP tunneling in RFC 1853 (RFC 1853, 1995). One protocol I’ll be using, called the Generic Routing Encapsulation (GRE) protocol, was developed by Cisco and others and standardized in October 1994, at around the same time (RFC 1701, 1994). Although tunneling protocols are a significantly newer concept than routing protocols, they developed from a similar origin and had a similar impact.

I’ll also be using a Virtual Private Network (VPN) protocol called Wireguard, which began development in 2016, written by Jason A. Donenfeld. It has quickly become one of the most popular VPN protocols due in part to its simplicity, which also led to it being integrated into the Linux kernel in January 2020 (Kernel/Git/Torvalds/Linux.Git - Linux Kernel Source Tree, n.d.), only growing its popularity. However, the idea for VPNs, or at least for encrypting data in transit, began in the early days of the ARPANET in the 1970s, when government agencies began looking for a way to establish secure communications over the new networking innovation. However, it wouldn't be until August 1995, when the Internet Engineering Task Force (IETF) convened with technology companies like Cisco to come up with a standard way to protect internet traffic, that IPsec would be developed (RFC 1826, 1995) (RFC 1827, 1995). When combined with the tunneling protocols mentioned above, this formed a secure way of tunneling traffic, what we now call "VPN protocols". A few years later, in 1999, Microsoft also released their own VPN standard, the Peer to Peer Tunneling Protocol (PPTP), which was specifically designed as a VPN and not to simply encrypt Internet Protocol traffic in transit, and used the GRE protocol for tunneling (RFC 2637, 1999). While other protocols like OpenVPN have been released since then, IPsec remains in widespread use despite its incredible complexity, vendor incompatibilities, poor performance, and constant vulnerabilities, and it’s only recently with Wireguard that many of these problems have been fully addressed. So fully, in fact, that the protocol was recommended by US Senator Ron Wyden to NIST as a potential replacement for the older protocols before it even left beta (Larabel, 2018).