Skip to main content

Lighthouse Protocol

version 0.1

Overview​

Lighthouse is the task dispatching protocol of the Kumulus Cloud.
It ensures that computational workloads are distributed optimally, securely, and fairly across multiple infrastructure tiers:

  • Spark — Mobile devices contributing light compute (on the edge)
  • Flare — PCs, laptops, and small servers contributing moderate compute
  • Stars — Data centers and enterprise-grade nodes providing heavy compute

Additionally, in the Stars tier, we have Community Nodes (C-Nodes) that represent regionally funded or community-owned data centers.

As Kumulus evolves, the Lighthouse protocol will serve as the core orchestrator that maintains fairness, transparency, and efficiency across this multi-tiered decentralized infrastructure.

Motivation​

In decentralized compute networks, resources vary in capability and reliability. Some devices are ephemeral (mobile), others persistent (data centers).
Lighthouse is designed to:

  • Ensure fair workload distribution across tiers.
  • Respect device constraints (compute power, bandwidth, energy).
  • Enable market-like matching between user demands and provider supply.
  • Maintain transparency and auditability in task assignment.
  • Support evolving product needs within the Kollectyve ecosystem.

Task Model​

At this stage, a task refers to any containerized workload that can be executed by a provider.

Examples include:

  • Spinning up a containerized VM.
  • Running a containerized AI inference service.
  • Executing a user-selected containerized app from a marketplace (e.g., wordpress, databases, web servers, language environments).
  • Serving as an edge compute node for partner protocols (like decentralized CDNs or AI training networks).

Essentially, a task can be any unit of compute that fits inside a containerized or virtualized environment

Revenue Model​

Kumulus business model builds on three pillars:

  1. Compute Marketplace
    Users (developers, businesses, institutions) pay for workloads run on the network.

  2. Provider Incentives
    Resource providers (Spark, Flare, Stars) earn rewards proportional to their contribution, uptime, and performance.

  3. Protocol Revenue
    A small portion of each transaction feeds back into protocol maintenance, ecosystem development, and community incentives.

This model aligns incentives between users, providers, and the network’s sustainability.

Lighthouse is not static — and it's design will evolve alongside Kumulus.
Early versions may use deterministic matching based on metrics like load, latency, and cost.
Later versions might incorporate machine learning and other tools for better workload dispatching.