Mesh Internet Networks

The original ARPANET design brief included a specific requirement: the network must continue to function even if significant portions of it are destroyed. This was 1969 and the context was nuclear war planning, but the engineering response to that requirement produced something more broadly useful — a packet-switched network that routes dynamically around failure. Every packet finds its own path. No single node is essential. The internet we have inherited from that design has been progressively re-centralized by commercial and governmental interests over the past thirty years, to the point where most traffic in any given country flows through a small number of choke points that are trivially controlled.

Mesh networking is the recovery of the original design. Not as nostalgia, but as engineering appropriate to the conditions communities actually face: ISPs that redline rural areas and low-income urban neighborhoods, governments that mandate surveillance backdoors, natural disasters that knock out backbone infrastructure, and corporate platforms that treat access as a product to be monetized rather than a commons to be governed.

The technical substrate of mesh networking has three layers worth understanding in detail.

At the radio layer, mesh networks use radio transceivers operating in unlicensed spectrum — primarily the 2.4 GHz and 5 GHz bands used by WiFi, but also 900 MHz for longer range with greater obstacle penetration, and sometimes licensed amateur radio bands for specific applications. Unlicensed spectrum means anyone can transmit within power limits set by the FCC or equivalent regulators without obtaining a license per installation. This is foundational to community mesh deployment; licensing requirements for each node would make volunteer-operated networks legally impossible. Hardware options range from consumer-grade equipment running modified firmware to purpose-built outdoor units from manufacturers including Ubiquiti, MikroTik, and TP-Link designed for point-to-point and point-to-multipoint links. The key performance parameters are transmit power, antenna gain, receiver sensitivity, and the modulation schemes supported. For urban mesh deployment, most nodes operate at ranges of 100-500 meters. For rural backbone links, directional antennas and clear line-of-sight allow reliable links of 5-20 kilometers.

At the routing layer, mesh networks use protocols designed to discover topology dynamically and route packets accordingly. The dominant open-source protocols are BATMAN-adv (Better Approach To Mobile Adhoc Networking — advanced), OLSR (Optimized Link State Routing), and Babel. Each has different tradeoffs in terms of convergence speed, overhead, and suitability for networks of different sizes. LibreMesh, an open-source firmware stack built on OpenWrt, integrates Babel and BMX7 and has become the standard platform for community mesh deployments globally. It includes a web interface for node configuration, automatic IP address assignment, and network visualization tools. Firmware installation on compatible hardware takes under 30 minutes for a competent technician.

At the application layer, a mesh network can carry any internet protocol traffic — web, email, voice over IP, video streaming — but it can also carry applications that function entirely within the local mesh without internet access. This is the capability that distinguishes community mesh from "cheaper broadband." A mesh-local application server — hosting a community forum, a file-sharing system, an emergency communications board, a local mapping tool — operates as long as any mesh node is running, regardless of internet connectivity. This is what disaster resilience actually looks like in network terms. When the Tohoku earthquake and tsunami struck Japan in 2011, local mesh networks in some communities maintained internal communications for days while backbone infrastructure was down. Similar patterns have appeared in Puerto Rico after Hurricane Maria, in Catalonia during political unrest, and in multiple conflict zones where external communications were interdicted.

The governance question for community mesh networks is not a soft organizational concern — it is technically load-bearing. Consider a few specific governance challenges. Gateway management: a mesh network with multiple gateway nodes connecting to commercial ISPs must manage whose ISP connection handles which traffic and how costs are allocated. If one member provides a gateway using their personal ISP account, they are legally responsible for all traffic passing through it. This creates liability exposure that must be addressed through either legal structure (an LLC or nonprofit holding the gateway connections) or technical measures (VPN tunnels that route gateway traffic through a jurisdiction-neutral endpoint). Node ownership: if a member installs a node on their roof and then moves, who owns the hardware? Who is responsible for maintaining it? Networks that have not resolved this in writing tend to degrade as founding members turn over. Acceptable use: a mesh network is a shared resource. Without enforceable policies on traffic types, bandwidth consumption, and prohibited uses, high-volume users or malicious actors can degrade service for all participants. These policies require both technical implementation (traffic shaping at gateway nodes) and social enforcement (membership agreements, processes for addressing violations).

The most sophisticated community mesh projects have resolved these questions through formal organizational structure. Guifi.net operates under a "Compact for a Free, Open and Neutral Network" — a legal agreement that participants sign, establishing mutual obligations for infrastructure sharing and traffic peering. The compact has been litigated in Spanish courts and upheld. NYC Mesh operates as a project of the nonprofit Internet Society New York Chapter, giving it legal standing to enter contracts, receive donations, and hold liability insurance. These structures allow the networks to grow beyond the capacity of purely informal volunteer coordination.

Financing models for mesh networks vary and each has distinct implications for sovereignty. Donation-based models (NYC Mesh) keep the network independent but create revenue uncertainty that limits long-term infrastructure investment. Member fee models (many European networks) create stable revenue but require enough members to cover costs, making early-stage networks financially fragile. Municipal partnership models — in which a city government provides grants, right-of-way access, or backbone infrastructure in exchange for coverage commitments — can accelerate growth but risk political dependency. The city of Barcelona funded significant portions of the Guifi.net backbone extension as part of its digital sovereignty agenda. The city of Detroit has supported mesh network projects in underserved neighborhoods as part of its broadband equity strategy. In both cases, the municipal relationship added capacity but also added complexity to network governance.

For communities planning a mesh deployment, the critical path has four phases. The first is technical feasibility: map the territory, identify high-elevation sites for backbone nodes, model coverage using radio propagation software, and test links with temporary equipment before committing to permanent installation. The second is governance design: establish the legal structure, draft membership agreements, define maintenance responsibilities, and set policies before the first node goes live. Retrofitting governance onto an operating network is far harder than building it in from the start. The third is phased deployment: start with a small connected cluster — three to five nodes minimum to demonstrate the mesh property — and expand methodically rather than deploying many isolated nodes that are not yet interconnected. The fourth is capability building: at least two or three community members need to develop genuine technical depth with the firmware, routing protocols, and radio hardware. Networks that depend on a single technical person are one departure away from collapse.

The political dimension of mesh networks is not incidental. In countries with sophisticated surveillance infrastructure, mesh networks that carry encrypted traffic represent a meaningful reduction in the state's visibility into community communications. In countries where internet shutdowns have been used as tools of political control — there were 182 documented intentional internet shutdowns in 2021 according to Access Now — mesh networks that can function as local communications infrastructure without internet access represent something qualitatively different from commercial resilience planning. This is not a call to paranoia. It is a call to honest assessment of what communications infrastructure controlled by others actually means for community sovereignty.

The technical barrier to mesh networking has dropped dramatically over the past decade. The firmware is mature, the hardware is affordable, the protocols are well-documented, and there is a global community of practitioners who have solved most of the practical problems that arise in deployment. What remains scarce is organizational will: the sustained collective effort to plan, deploy, govern, and maintain community infrastructure rather than paying a corporation to do it on terms that serve corporate interests. That scarcity is not a technical problem.