This preliminary report is organized into two main sections: the first introduces the original Cryptarchia v2 protocol, outlining its core concepts and presenting a structured roadmap for its security analysis. The second section focuses on Cryptarchia v2-minus, a deliberately simplified variant designed to facilitate rigorous security evaluation. By examining v2-minus from multiple perspectives, the report identifies and analyzes various potential attack vectors—including those already recognized in the original v2 proposal—and then introduces a modified security model that considers a purposely restricted class of adversaries.

Cryptarchia v2

1. Summary of the Cryptarchia v2 Protocol

Cryptarchia v2 is a proof-of-stake (PoS) consensus protocol that combines new mechanisms with ideas from existing literature to improve resistance to chain reorganizations, better handle network delays, and enhance censorship resistance through support for multiple concurrent proposers. The protocol’s core structure is a directed acyclic graph (DAG), where each block includes independent transactions and references to other blocks within a sliding window. These references form a maximal antichain—ensuring no referenced block is reachable from another—which promotes independent voting and accelerates convergence.

The fork choice rule selects the branch whose closest common ancestor (CCA) has the highest cumulative weight, where weight is defined as the number of window-referenced blocks in its descendants. An optional “long-ref” allows a block to reference one block outside the window for improved DAG connectivity, but it does not affect fork choice.

1.1 Core Mechanics

• Multi-Proposer DAG: All registered validators (executors) can produce blocks in parallel each slot. Blocks include a random subset of UTXO transactions (to resist tagging attacks), references to other blocks (forming the DAG), and dependency edges to encode UTXO consumption and prevent double spends.
• Reference Window and Antichain: Each block references as many recent blocks as possible (default window: 30 slots), forming a maximal antichain. An optional long-ref can point to an older block for connectivity but does not impact conflict resolution.
• Conflict Resolution: For conflicting transactions, nodes identify the CCA of the conflicting blocks and compare the cumulative weight of descendant blocks within the window. The heavier branch is chosen.
• Slashing: Short-range equivocations, such as double-signing, are punished via on-chain slashing transactions.
• Experimental Results: Simulations with up to 49% adversarial stake show improved stability, faster convergence, and better tolerance to delays compared to v1. The design supports faster finality through increased block production, although more analysis is needed.
• Unpredictability: The current version lacks a finalized unpredictability mechanism for block eligibility. Future versions aim to incorporate this component.

1.2 Strengths

• High Throughput and Censorship Resistance: Parallel block production by all validators increases throughput and impedes censorship.
• Rapid Convergence and Network Adaptivity: Maximal antichain references and a sliding window improve convergence and adapt well to delays.
• Clear Fork Choice Rule: The CCA-based weight comparison generalizes established ideas (e.g., PHANTOM, Praos) into a deterministic rule.