Reticulum

From P2P Foundation Wiki
Jump to navigation Jump to search

= "Reticulum Network Stack is a protocol stack from Mark Qvist to create privacy-focused networks that can operate over LoRa, Ham Radio etc." [1]

URL =


Description

Synthesis by ChatGPT, prompted by Michel Bauwens, with emphasis on the ideas expressed in the essay, The Zen of Reticulum:

    • Reticulum** is a cryptography-based networking stack for constructing local and wide-area peer-to-peer networks over heterogeneous communication media. It is designed to function over links with very low bandwidth and high latency, including LoRa, packet radio, Wi-Fi, serial links and IP networks. Rather than constituting a single network, Reticulum is intended as a technology for building many autonomous and interoperable networks.

The system differs from conventional IP networking in several fundamental respects. It does not organise communication around the familiar combination of IP addresses, ports and centrally administered address spaces. Applications create **destinations**, represented cryptographically by destination hashes. These destinations are portable: they identify an endpoint through cryptographic identity rather than fixing it to a particular physical or network location. Reticulum can therefore discover paths to an endpoint as it moves between different parts and media of a network.

Reticulum also treats encryption and cryptographic authentication as intrinsic properties of the networking layer rather than optional additions. Its design incorporates end-to-end encryption, forward secrecy, cryptographic identity, authenticated links and multi-hop transport. Intermediate transport nodes can forward traffic without requiring a central directory or possessing the complete topology of the network.

The project's stated objective is consequently not simply to produce another decentralised networking protocol, but to make it possible for individuals and communities to operate autonomous communication infrastructure. In the project's own formulation, Reticulum is “not one network” but a tool for constructing potentially large numbers of independent networks that can interconnect without a single authority governing the whole.

Discussion

The Zen of Reticulum

The *Zen of Reticulum* presents the technical architecture of Reticulum through a broader critique of the dominant organisation of digital networks. Its argument can be read as a series of linked theses concerning **centralisation, identity, trust, infrastructure and autonomy**. The three concepts highlighted in the P2P Foundation Wiki — *Uncentralizability*, the *Abolition of the Digital Address*, and *Paranoia as a Design Principle* — are particularly useful for reconstructing this argument.


From decentralisation to uncentralizability

A first thesis is that **decentralisation is insufficient if the possibility of recentralisation remains structurally built into the system**.

The *Zen* distinguishes between systems that distribute functions among several centres and systems designed so that no centre is required in the first place. A federation, for example, may distribute authority among servers while leaving those servers as privileged points of coordination. Likewise, a nominally distributed system can acquire effective centres through concentration of infrastructure, administration or economic power.

Reticulum therefore uses the stronger concept of **uncentralizability**. The issue is not simply where authority happens to be located at a particular moment, but whether the architecture creates privileged positions from which the network as a whole can be controlled.

This is reflected technically in the absence of a centrally administered address space and in the absence of intrinsically privileged routing nodes. Transport nodes can forward traffic, but they are not conceived as authorities over the identities or communications that pass through them. The P2P Foundation's *Uncentralizability* entry consequently interprets Reticulum as an attempt to make hierarchy a property that the protocol itself resists rather than merely a political preference of its operators.

The underlying proposition can therefore be stated as:

> **A peer-to-peer system is more robust against centralisation when its essential functions do not require a privileged centre to exist.**


From the “cloud” to a network of peers

A related thesis is a rejection of the familiar mental model of **“the cloud”**.

The *Zen* argues that what is ordinarily presented as an abstract cloud is in practice a collection of computers, infrastructures and institutions operated by particular parties. The conventional client-server model makes this asymmetry almost invisible: users are positioned as clients who connect to servers that possess resources, authority and continuity.

Reticulum proposes a different conceptual starting point: rather than asking how a client connects to a service “up there”, one asks how autonomous participants can exist **among one another**.

This distinction is significant for P2P theory because it shifts decentralisation from an organisational property of a service to an architectural property of the network. The objective is not merely to distribute the servers of an existing service, but to make it possible for participants themselves to provide infrastructure and services.


The abolition of the digital address

A central architectural thesis is that **identity should not be synonymous with location**.

Traditional network addresses perform a fundamentally geographical or topological function: they indicate where a machine can be found within a particular addressing scheme. Consequently, movement between networks can require a change of address, while an address can reveal information about the location or network attachment of an endpoint.

Reticulum separates these functions. Its destinations are derived from cryptographic identities rather than from a centrally allocated numerical location. A destination can therefore remain associated with an identity while the physical or network location through which it is currently reachable changes.


The network should route towards identities, not locations

The previous thesis leads to a further distinction between **routing** and **addressing**.

In conventional networking, routing is generally understood as the process of finding a path through a hierarchy of network locations. Reticulum instead separates the cryptographic identity of a destination from the changing paths by which it can be reached.

The network can consequently discover paths towards destinations without requiring a global administrative map of all participants. Transport nodes maintain enough local knowledge to forward traffic while the network converges on reachable destinations.

This is important because it removes one of the classic sources of centralisation: the institution that maintains the authoritative directory of who exists and where they are.

The architectural principle is therefore:

> **Global reachability need not imply global administration.**


Paranoia as a design principle

The *Zen* then applies a particular security philosophy to the resulting peer-to-peer environment: **assume that the infrastructure cannot be trusted**.

The P2P Foundation's *Paranoia as a Design Principle* summarises this as an inversion of the conventional “trusted core” model. Instead of assuming that local networks, backbone providers, routers or other infrastructure operators are benign, Reticulum treats links and intermediary nodes as potentially compromised.

This is not “paranoia” in the psychological sense. It is a design assumption:

    • if an intermediary must be trusted for the system to remain secure, that intermediary constitutes a potential point of failure.**

Reticulum therefore seeks to transfer trust away from infrastructure and towards cryptographic properties of identities, messages and links. A node is not trusted merely because it belongs to a recognised institution or occupies a privileged place in the network. Cryptographic verification supplies the basis for establishing the relevant relationships.


Encryption as infrastructure rather than an add-on

This produces another important thesis: **security should not be a supplementary layer applied to an otherwise trusting network.**

In many conventional architectures, encryption is presented as a service layered over networking — for example, through TLS or VPNs. Reticulum instead incorporates cryptographic mechanisms into the fundamental operation of communication and transport. Its documentation describes encryption, authentication and forward secrecy as integral characteristics of the stack.

The philosophical implication is that privacy is not treated as an exceptional requirement for particularly sensitive communications. The network begins from the assumption that communication should not depend upon trusting every intermediary through which it passes.

Thus:

> **The infrastructure is permitted to be untrusted because the communication does not require the infrastructure itself to be trusted.**


From institutional trust to cryptographic verification

The *Zen* consequently proposes a broader transformation in the meaning of **trust**.

Traditional digital systems frequently establish trust through institutions: certificate authorities, domain registries, telecommunications providers, platform operators, application stores, server operators and other administrative intermediaries. Reticulum attempts to replace much of this institutional mediation with direct cryptographic verification.

This does not mean that social trust disappears. Rather, the architecture attempts to distinguish **trust in a person or service** from **trust in the infrastructure carrying the communication**.

A participant may decide whom to recognise, while the protocol provides mathematical means for verifying that the entity communicating under a given cryptographic identity actually possesses the corresponding credentials.

The result is a redistribution of agency:

    • institutional model:**

authority → certifies identity → user trusts authority


    • cryptographic model:**

identity → proves possession of key → participant verifies identity


A synthetic statement of the Reticulum thesis

Taken together, the arguments can be condensed into a single proposition:

    • Digital networks should be designed so that autonomy does not depend upon the continued benevolence, availability or permission of a central infrastructure.**


From this proposition follow several architectural consequences:

1. **No necessary centre:** essential network functions should not require a privileged authority.

2. **Portable identity:** identity should be separable from physical location.

3. **No central address authority:** participants should be able to create and control their own destinations.

4. **Distributed routing:** global connectivity should emerge from local cooperation rather than from a single global directory.

5. **Untrusted infrastructure:** intermediary networks should be assumed potentially hostile or compromised.

6. **Cryptographic trust:** authentication should rest on verifiable cryptographic properties rather than institutional privilege.

7. **Encryption by default:** confidentiality and integrity should be inherent to communication rather than exceptional features.

8. **Infrastructure sovereignty:** individuals and communities should be able to operate their own network segments and interconnect them voluntarily.

9. **Heterogeneous networking:** different physical media and scales should be capable of participating in the same logical communication fabric.

10. **Resilience through multiplicity:** the network should consist of many independently operated networks rather than a single infrastructure whose failure or control compromises everyone.


In this sense, **Reticulum can be understood as an attempt to translate a P2P political and architectural intuition into networking primitives**. Its distinctive claim is not simply that networks should have “more peers” or “fewer servers”, but that the protocol should make autonomy, mobility, cryptographic identity and distributed infrastructure mutually reinforcing.

The *Zen of Reticulum* thus moves from a critique of the **centre**, through a critique of the **address**, to a critique of **trust**. The centre is displaced by a fabric of peers; the fixed address is displaced by portable cryptographic identity; and institutional trust is displaced, where possible, by cryptographic verification. The resulting model is one in which the network's continuity depends less upon a particular institution or infrastructure than upon the continued existence of relationships among autonomous participants.


More information

Excerpts from the 'Zen of Reticulum' via:

  1. Uncentralizability
  2. Abolition of the Digital Address
  3. Paranoia as a Design Principle