KRILL: Bio-Inspired Network Architecture for the Internet of Things.

Abstract

KRILL — Bio-Inspired Architecture for IoT Consensus Decentralized IoT consensus without blockchain — inspired by ant colonies, immune systems & chemical diffusion. What is KRILL? The problem: Blockchain doesn't work for IoT. It's too heavy, too slow, and too expensive for devices running on batteries with 32KB of RAM. IoT needs to answer "What is the physical state of the world?" — not "Who has how much money?" The solution: KRILL replaces blockchain with 9 mechanisms borrowed from biology: Mechanism Biological inspiration What it does Stigmergic Consensus Ant pheromone trails Nodes "deposit" readings like ants deposit pheromones. Truth emerges from convergence, not voting. Pentastratic Immune System Human immune layers 5-layer anomaly detection: skin (format check) → innate (statistical) → adaptive (learned) → NK audit → autoimmune suppression. Metabolic State Cell metabolism Data has a "half-life" — old readings decay and die automatically. No infinite ledger. Entropic Data Valuation Thermodynamic entropy Network autonomously decides which data is worth storing based on information theory. Quorum Sensing Bacterial quorum sensing Nodes detect local density and switch modes (solo → quorum → swarm) without any coordinator. Horizontal Gene Transfer Bacterial gene sharing Firmware updates spread node-to-node like genes between bacteria. No update server needed. Morphogenetic Topology Embryonic development Network self-organizes its topology using reaction-diffusion (Turing patterns). Thymic Tolerance T-cell training in thymus System learns what "normal" looks like to avoid false alarms. Immunological Memory Vaccine/antibody memory Once the network detects an attack pattern, it "vaccinates" all nodes. The result: 1000x less energy than blockchain consensus Runs on a $2 ESP32 microcontroller (240KB RAM) Works with intermittent connectivity (mesh, BLE, LoRa, WiFi) No miners, no staking, no tokens — consensus is grounded in physical reality Scales to millions of nodes without coordinator Status: Research paper + engineering specification. No working implementation yet. Documents Document Description Research Paper (HTML) Full academic paper — mathematical formalizations, energy analysis, novelty assessment, risk analysis. 20 sections. Open in browser → Print → Save as PDF. Engineering Specification (HTML) Implementation reference — byte-level wire formats, state machines, pseudocode, test vectors, transport layers. Ready to code from. Source files (Markdown): krill-bioinspired-architecture.md — Research paper krill-bia-engineering-spec.md — Engineering spec Architecture at a Glance ┌─────────────────────────────────────────────────────────┐ │ KRILL Node (ESP32) │ ├──────────┬──────────┬──────────┬──────────┬─────────────┤ │ Stigmer- │ Immune │ Metabolic│ Quorum │ Morpho- │ │ gic │ System │ State │ Sensing │ genetic │ │ Consensus│ (5-layer)│ (decay) │ (modes) │ Topology │ ├──────────┴──────────┴──────────┴──────────┴─────────────┤ │ Transport: BLE mesh / WiFi / LoRa │ ├─────────────────────────────────────────────────────────┤ │ PUF Identity + Ed25519 Enrollment │ └─────────────────────────────────────────────────────────┘ MVP — Where to Start If you want to implement KRILL, start with these 4 subsystems (the rest can be added later): ES-13 — Cryptographic enrollment (PUF + Ed25519 identity) ES-12 — Transport layer (BLE mesh for local, WiFi for bridging) ES-1 — Core data types and wire formats ES-3 — Stigmergic Consensus (the core algorithm) ES-10 — Main event loop and message dispatch Target hardware: ESP32 (Nano node) + nRF52840 (Dust node, optional) Why Not Blockchain? Blockchain (e.g. Ethereum) KRILL-BIA Consensus energy ~50 Wh/tx (PoW) or ~0.01 Wh/tx (PoS) ~0.00001 Wh/tx Minimum RAM 512MB+ 32KB (Dust), 240KB (Nano) State growth Infinite (append-only) Bounded (data decays) Offline tolerance Minutes before fork Days (pheromone half-life) Hardware cost $50+ SBC $2 ESP32 Finality Probabilistic (blocks) Convergent (pheromone field) Key Innovation: Physical-World Consensus Grounding Unlike blockchain where consensus is purely computational, KRILL grounds consensus in physical reality: Sensor readings must be physically plausible (a thermometer can't jump 50C in 1 second) Nodes that are physically closer have more weight (radio signal strength = distance proxy) The laws of physics constrain what values are possible — this is a defense layer that doesn't exist in financial systems This means an attacker must not only compromise the software but also defeat physics — a fundamentally harder problem. Contributing See CONTRIBUTING.md for how to get involved. Areas where help is most needed: Rust/C firmware for ESP32 (core protocol implementation) Simulation — model pheromone convergence with 100-10,000 virtual nodes Hardware testing — BLE mesh range, LoRa timing, PUF enrollment on real chips Security review — formal verification of immune system thresholds Documentation — diagrams, tutorials, translations License This project is licensed under the MIT License. Supporting This Work If KRILL is useful to your research or organization, consider supporting further development: ETH / ERC-20 / Base / Arbitrum / Polygon: 0x0BC290355c0B16B5B247701B7BC9AB2E1e61ffa7 Funds go toward: Reference firmware for ESP32 + nRF52840 Hardware test beds (100-node BLE mesh) Independent security audits Bug bounty program for protocol vulnerabilities Code contributions are equally welcome — see CONTRIBUTING.md.