Cardano PubSubR&D

A verifiable communication layer for the Cardano ecosystem

A Cardano-anchored publish/subscribe layer for authenticated ecosystem communication: identifiable publishers, verifiable events, on-chain registered topics, and an economic cost to creating publisher identities.

Why it matters

Critical communication deserves the same guarantees as transactions

Node operations

Node builder to operator

Node builders — teams like IOG, Amaru, and the Cardano Foundation — reach node operators with authenticated emergency and operational alerts.

Stake pools

SPO to delegator

Stake pool operators announce pool changes, incidents and fee updates to their delegators, through a topic that provably belongs to the pool.

Governance

dRep to delegator

dReps communicate voting intent and rationale to the delegators they represent under CIP-1694.

Applications

dApp to user

dApps announce critical alerts and per-user notifications under on-chain identities, with spam resisted by real stake.

Your turn

Have a use case we have missed?

Tell us who needs to reach whom, and what stands in the way today. It opens a discussion on GitHub, where anyone can read it, add to it, or back it.

Research programme

From proposed design to evidence-based architecture

Jul 2023
Foundations

SecureCyclon published

Antonov & Voulgaris publish SecureCyclon at ICDCS 2023 — a peer-sampling protocol with defences against protocol-violating peers, proposed as the decentralised foundation the PubSub workstream would build on.

Feb 2026
Phase 1 · Complete

Exploratory research

Use-case and requirements analysis, cryptographic key-material study, and formal plus empirical analysis of the proposed peer-sampling foundation. The analysis surfaced silent attack vectors — adversaries can bias nodes' views of the network without triggering any of the protocol's defences — which led to the sampling layer being set aside in favour of an on-chain, list-based design.

Jul 2026
Phase 2 · Active

Empirically-driven architecture

Five candidate dissemination models, from simple push to hybrid designs, analysed formally and measured empirically against a silent Byzantine adversary — so the final design is the one the data supports, not the one intuition favours. A Rust prototype node implements topics, subscriptions, publishing, and fan-out forwarding, backed by a deterministic simulation framework for controlled experiments.

Ahead
Planned

Hardening & ecosystem integration

A design proposal consolidates the analytical and experimental findings into a concrete architecture recommendation — published as a CIP if phase 2 time permits. The scope beyond that is still taking shape; candidates include a fee-and-incentive model that rewards honest participation, real network transport in the prototype node, a public testnet, and — on the research side — decentralised peer sampling at scales where the on-chain list no longer suffices.

Experiments

Five candidate models, one bar

Not proposing a design, but empirically deriving one. All five face the same bar: every honest publisher's message reaches every honest node despite silent Byzantine peers, measured as the probability that the honest network stays strongly connected. All five can clear it. What separates them is the price: how many standing links every node has to hold, and how many copies of each published message the network has to carry.

Development activity

Contributions over the last 6 months

Research logbook

The running record of decisions, experiments, and milestones as they happen.