TACON-Sui Technical Paper
Optimizing Sui's Network Layer: A Topology-Aware and Cost-Optimized Approach for Consensus and Data Dissemination
Sui blockchain has distinguished itself through its object-centric data model, parallel transaction execution, and the high-performance Mysticeti consensus protocol. However, the efficiency of any distributed consensus mechanism is intrinsically linked to the underlying peer-to-peer (P2P) networking layer responsible for block propagation, synchronization, and state dissemination.
Currently, Sui's networking layer operates without traditional gossip protocols between validators for DAG block dissemination, instead employing a "direct push" mechanism for newly created blocks and "indirect polling" for missing data. While functional under ideal conditions, this approach lacks awareness of underlying network topology and varying communication costs, potentially leading to suboptimal routing and reduced resilience.
This paper proposes TACON-Sui (Topology-Aware Consensus Networking for Sui), a comprehensive solution designed to enhance the networking aspects of Sui's consensus and critical P2P components. TACON-Sui introduces intelligent networking subsystems that actively discover network topology, estimate path costs, and utilize this intelligence to optimize data propagation across all phases: push, pull (sync), and checkpoint dissemination.
Mysticeti Consensus Protocol
Sui's current consensus protocol, Mysticeti, is a DAG-based Byzantine Fault Tolerant (BFT) protocol designed for low latency and high throughput. Key benefits include parallel block proposals, commitment in three message rounds, and tolerance for unavailable leaders.
Performance Metrics
- • 200,000-400,000 TPS in controlled tests
- • Sub-second latency capability
- • 390-500ms finality observed
Architecture Features
- • Single message type (signed block)
- • Single multi-cast transmission
- • Implicit commitment mechanism
Current Network Limitations
Direct Push Mechanism
Validators directly push newly created DAG blocks to all other validators without considering network topology or path costs.
Indirect Polling Strategy
Missing blocks are detected and requested through polling, but without intelligent peer selection based on network proximity or reliability.
No Topology Awareness
The system lacks global network view and cannot adapt to changing network conditions or optimize for different communication costs.
Core Concepts
Dynamic Network Topology Discovery
Continuously build and update a map of the validator and relevant full node network, including their interconnections and geographic locations.
Path Cost and Quality Estimation
Actively and passively measure various metrics (latency, bandwidth, jitter, reliability) for paths between nodes.
Intelligent Peer Selection
Utilize topology and cost data to select optimal peers for pushing blocks, pulling missing data, and synchronizing checkpoints.
Optimized Data Propagation
Employ distinct strategies for direct push (multicast/broadcast optimization) and indirect pull (targeted requests).
System Components
Network Topology and Cost Service (NTCS)
Central service responsible for discovering, maintaining, and providing network topology information with associated cost metrics.
Topology-Aware Routing Module (TARM)
Handles DAG block propagation with intelligent peer selection and multicast optimization strategies.
Smart Checkpoint Dissemination Protocol (SCDP)
Optimizes checkpoint summary gossip and bulk data transfer using parallel streams and erasure coding.
SCION Integration Layer
Leverages SCION's next-generation internet architecture for verified, policy-compliant path selection.
Quantitative Improvements
Latency Reduction
Reduced tail latencies through optimal path selection
Intelligent pull requests to responsive peers
Throughput Enhancement
Optimal path selection maximizes available bandwidth
Parallel streams and erasure coding
System Resilience Improvements
Network Fault Tolerance
Adaptive routing around failed paths and congested links with automatic failover.
Geographic Distribution
Optimized performance across globally distributed validator sets with regional preferences
DDoS Mitigation
Enhanced protection against network-layer attacks through intelligent path diversity.
Development Phases
Phase 1: Foundation (Q1 2025)
Core NTCS development and Anemo integration layer implementation.
- • Basic topology discovery mechanisms
- • Cost modeling framework
- • Anemo RPC interface design
Phase 2: Core Modules (Q2 2025)
TARM and SCDP implementation with intelligent peer selection algorithms.
- • DAG block propagation optimization
- • Checkpoint dissemination enhancement
- • Performance monitoring integration
Phase 3: SCION Integration (Q3 2025)
Advanced path selection with SCION network architecture integration.
- • SCION path discovery and selection
- • Policy enforcement mechanisms
- • Geo-fencing and AS preference support
Phase 4: Production Deployment (Q4 2025)
Mainnet integration and enterprise-grade monitoring capabilities.
- • Mainnet validator deployment
- • Enterprise integration tools
- • Community adoption programs
Future Research Directions
Machine Learning Integration
Adaptive algorithms for dynamic network optimization and predictive path selection based on historical performance data.
Formal Verification
Mathematical verification of TACON-Sui protocols to ensure correctness and resilience properties.
Cross-Chain Optimization
Extension of topology-aware networking to multi-chain environments and interoperability protocols.
Quantum-Resistant Security
Integration of post-quantum cryptographic methods for long-term security in network communications.
- Introduction
- Sui's Current Network Architecture and Identified Challenges
- Consensus: Mysticeti and its Predecessors
- DAG Block Propagation: Direct Push and Indirect Polling
- Checkpoint Dissemination and State Synchronization
- Anemo P2P Library Integration
- TACON-Sui: A Topology-Aware and Cost-Optimized Networking Framework
- Core Concepts and Design Principles
- Network Topology and Cost Service (NTCS)
- Topology-Aware Routing Module (TARM)
- Smart Checkpoint Dissemination Protocol (SCDP)
- SCION Integration for Advanced Path Selection
- Implementation Architecture and API Design
- System Architecture Overview
- Anemo RPC Interface Extensions
- Cost Metrics and Measurement Framework
- Dynamic Configuration and Policy Management
- Anticipated Performance Gains and System Resilience
- Quantitative Improvements
- Latency and Throughput Analysis
- Network Resilience Enhancements
- Scalability Considerations
- Enterprise Deployment and Integration
- Deployment Strategies
- Monitoring and Analytics
- Security Considerations
- Compliance and Governance
- Conclusion and Future Directions
- References and Citations