Mondoshawan Protocol

0 Abstract

Mondoshawan is a Layer 1 blockchain protocol built in Rust that combines three capabilities no existing production chain currently offers simultaneously: GhostDAG BlockDAG consensus for parallel block production, ML-KEM-768 post-quantum cryptography (NIST FIPS 203) on all peer-to-peer transport, and full EVM compatibility via SputnikVM. The protocol is currently operational on a multi-datacenter public testnet, producing blocks at a stable ~10-second median interval using real cryptographic Proof of Work.

This document describes the technical architecture of the Mondoshawan protocol as it exists in the current codebase, distinguishing implemented features from planned work. Claims are grounded in the live implementation.

1 Motivation & Problem Statement

The Quantum Threat is Not Hypothetical

In August 2024, NIST finalized three post-quantum cryptographic standards: FIPS 203 (ML-KEM, key encapsulation), FIPS 204 (ML-DSA, digital signatures), and FIPS 205 (SLH-DSA, hash-based signatures). The US National Security Agency has mandated migration to these standards for all classified systems by 2030. The Federal Reserve has published guidance on the "Harvest Now, Decrypt Later" (HNDL) threat — the practice of recording encrypted network traffic today for decryption once a sufficiently powerful quantum computer exists.

The Gap in the Market

Examining the current landscape reveals a clear gap. Kaspa implements GhostDAG with high throughput but has no post-quantum security and no EVM. Ethereum has EVM and a large developer ecosystem but uses a single-chain architecture and has no post-quantum roadmap at the protocol layer. QRL implements post-quantum signatures but has no EVM and uses a traditional single chain. No production Layer 1 combines all three.

Mondoshawan is built to occupy this intersection.

2 System Architecture

The Mondoshawan node is structured as four distinct layers, each with clear responsibilities and interfaces:

3 GhostDAG Consensus

BlockDAG vs. Single Chain

Traditional blockchains discard blocks produced simultaneously (orphans), wasting the honest work of miners and limiting throughput to a single block per interval. GhostDAG (Greedy Heaviest-Observed Sub-DAG) resolves this by organizing all produced blocks into a Directed Acyclic Graph and then computing a linear ordering over them.

The protocol uses a parameter K (set to 4 in Mondoshawan's testnet configuration) representing the maximum number of blocks produced in a single network propagation delay window. Blocks are classified as blue (honest, well-connected) or red (potentially dishonest or poorly connected) based on their position in the DAG relative to the blue set. Only blue blocks earn mining rewards.

Storage Architecture

The DAG is stored using a hybrid strategy: a sled embedded key-value store provides durable persistence, while a bounded in-memory LRU cache holds the "hot" tip region of the DAG for low-latency blue score computation. This avoids the recursive traversal penalty that would otherwise make GhostDAG impractical on a key-value store.

Sync Resilience

The sync layer implements compact block IBD (analogous to BIP152), stall detection with automatic re-sync triggers, and quorum-based fork recovery. These mechanisms were hardened through extensive multi-datacenter testing and are running in production on the current testnet.

4 BraidCore Mining Architecture

Mondoshawan introduces a three-stream parallel mining architecture where each stream uses a different algorithm targeting different hardware profiles. All three streams feed blocks into the same GhostDAG consensus engine.

Economic Rationale

The BraidCore model addresses a fundamental tension in PoW design. Pure ASIC mining (Bitcoin) concentrates hash power in large industrial operations. Pure memory-hard mining (Monero) resists ASICs but limits throughput. By running both in parallel and adding a ZK proof validation stream, Mondoshawan creates three distinct economic constituencies — ASIC operators, GPU miners, and ZK prover operators — each with different cost structures and incentives, making coordinated attacks significantly more expensive.

5 Proof of Work & Difficulty Adjustment

B3MemHash Algorithm

Stream B uses a custom memory-hard hash function called B3MemHash. The algorithm initializes a 256-byte scratchpad from the block header using Blake3, then performs 64 rounds of XOR mixing across the buffer using SIMD intrinsics (AVX2 on x86-64, SSE2 fallback, scalar on other architectures). The final hash is produced by running Blake3 over the mixed buffer.

Difficulty Adjustment Algorithm (DAA)

The DAA uses fixed-point arithmetic (no floating-point in consensus code) to compute the ratio of target block time to actual block time, then clamps the adjustment to a range of 0.667× to 1.5× per adjustment window. This prevents the oscillation behavior common in naive DAA implementations.

Two variants are implemented: a single-block adjustment for rapid response and a moving-average variant for stability over longer windows. The testnet currently runs with the following parameters:

6 Post-Quantum Cryptography

NIST Standards Implemented

Mondoshawan implements post-quantum cryptography at the network transport layer using ML-KEM-768 (Module Lattice Key Encapsulation Mechanism, NIST FIPS 203), the standardized version of CRYSTALS-Kyber. This provides 192-bit classical security and is resistant to Grover's algorithm and Shor's algorithm attacks.

The Kyber Handshake Protocol

Every peer connection performs a full ML-KEM-768 key exchange before any block data is transmitted. The handshake proceeds as follows:

Why This Matters: HNDL Protection

The Harvest Now, Decrypt Later attack allows a quantum-capable adversary to record classical ECDH-encrypted P2P traffic today and decrypt it once a sufficiently powerful quantum computer exists. By using ML-KEM-768 for key exchange, Mondoshawan ensures that even if all historical P2P traffic is archived, it cannot be decrypted retroactively — the shared secret is computationally infeasible to derive from the public encapsulation key alone, even with a quantum computer.

7 P2P Network Layer

The Mondoshawan P2P layer uses QUIC as the transport protocol, providing multiplexed streams, built-in TLS 1.3, and connection migration. Peer identity is established via TOFU (Trust On First Use) — each node presents a self-signed Ed25519 certificate, and the fingerprint is stored on first connection.

Sybil & Eclipse Attack Resistance

The network layer implements several mitigations against Sybil and eclipse attacks:

8 EVM Integration

Mondoshawan integrates SputnikVM at the Shanghai hard fork level, providing full EVM opcode compatibility. The JSON-RPC surface has been validated with 129+ methods covering the full eth_*, net_*, web3_*, and mds_* (Mondoshawan-specific) namespaces.

Validated Capabilities

Account Abstraction

The protocol implements native account abstraction with smart contract wallets supporting multi-signature authorization, social recovery mechanisms, per-transaction spending limits, and atomic batch transactions. This is implemented at the protocol level rather than as an L2 overlay, providing lower gas costs and tighter security guarantees than ERC-4337 on Ethereum.

9 Sharding

Mondoshawan implements native state sharding using consistent hashing for deterministic shard assignment. Each shard maintains its own state tree and transaction pool. Cross-shard transactions use a two-phase commit protocol with atomic finality guarantees.

10 Tokenomics

Stream A (Blake3) rewards 50 QMON per block. Stream B (B3MemHash) rewards 25 QMON per block. Stream C (ZK Proofs, planned) will be fee-based only. There is no pre-mine, no presale, and no VC allocation. The supply is entirely emission-based.

A portion of transaction fees are burned on each block, creating deflationary pressure as network usage grows. The burn rate and emission schedule are subject to governance adjustment post-mainnet.

11 Competitive Analysis

12 Roadmap

13 Conclusion

Mondoshawan is a production-grade Layer 1 blockchain protocol that addresses a clear gap in the current landscape. The combination of GhostDAG consensus, ML-KEM-768 post-quantum transport, and full EVM compatibility is unique among live or testnet-stage protocols as of April 2026.

The protocol is not a whitepaper or a proof of concept. It is a running system producing blocks at stable intervals between geographically distributed nodes using real cryptographic Proof of Work and real post-quantum key exchange. The core implementation is available for technical review at the GitHub repository below.