Introduction
Huawei Huang, Yetong Zhao, and Zibin Zheng* presented their groundbreaking research on blockchain sharding technology at the IEEE/ACM International Symposium on Quality of Service (IWQoS) in June 2023. Their paper, titled "tMPT: Reconfiguration across Blockchain Shards via Trimmed Merkle Patricia Trie," introduces an innovative approach to enhancing blockchain scalability through efficient shard reconfiguration.
Research Background and Motivation
Sharding technology has emerged as a promising solution for improving blockchain scalability [1-4]. By dividing consensus nodes into multiple shards, this technique enables parallel transaction processing, significantly boosting network throughput—a critical feature for high-transaction-rate platforms.
However, sharding introduces security challenges as the system's safety becomes distributed across individual shards rather than the entire network. Current solutions like Elastico [1], RapidChain [2], and Omniledger [3] propose periodic node reshuffling and partial synchronization methods, but these approaches remain in early developmental stages. Our research addresses this gap by presenting a comprehensive shard reconfiguration protocol that maintains security while improving efficiency.
Key Contributions
- Efficient Shard Reconfiguration Protocol: Our solution maintains system security while optimizing reorganization efficiency.
- tMPT Data Structure: The trimmed Merkle Patricia Trie compresses state trees within shards, significantly reducing synchronization overhead.
- Partial Reconfiguration: We introduce an inter-shard partial reorganization method to minimize system impact during reconfiguration.
- Performance Gains: Experimental results show our protocol achieves 198% higher throughput than Ethereum's Full Synchronization method.
Protocol Overview
Core Concept
Inspired by Ethanos [5] findings about Ethereum's transaction temporal locality—where certain accounts show high activity within short periods—our protocol synchronizes only active account states during reconfiguration. This selective approach dramatically reduces data transfer requirements.
System Architecture
Node Roles
Validator Nodes:
- Store historical block data and account states
- Participate in transaction validation
- Organized into multiple validator shards
Witness Nodes:
- Maintain complete network state
- Generate reorganization plans
- Assist in transaction validation
- Form witness shards
Workflow
👉 Discover how blockchain sharding transforms scalability
- Consensus Phase: Validator shards process transactions and create blocks through intra-shard consensus.
Reconfiguration Phase:
- Witness shards generate reorganization plans using VRF random functions
- Plans are broadcast network-wide after block finalization
- Validator shards trim state trees, removing nodes with access cycles < k
- Compressed state trees (epoch k) are transmitted to relevant nodes
- Nodes merge state information and reach consensus on updated states
Experimental Results
Setup
We implemented the tMPT protocol using Golang on our blockEmulator simulator (soon to be open-source). Tests used 1.5 million Ethereum transactions from July 2018, with:
- Block size: 1,000 transactions
- Block interval: 4 seconds
- 4 validator shards + 1 witness shard
- 4 consensus nodes per shard
- Deployment on Alibaba Cloud servers
Key Findings
- Throughput: tMPT methods achieved 3-3.4× higher TPS than Ethereum full sync
- Consistency: Maintained stable TPS while baseline methods showed declining performance
Efficiency:
- Final epoch reconfiguration latency: 2.8% of Ethereum full sync
- Data transfer: 13% of Ethereum full sync
Conclusion
Our tMPT-based shard reconfiguration solution significantly improves blockchain scalability while maintaining security. The protocol's efficiency gains in throughput and latency demonstrate its superiority over existing methods, making it a compelling approach for next-generation blockchain systems.
Industrial Application Prospects
While sharding technology remains in development, our tMPT protocol addresses critical challenges in shard reorganization—offering a secure, efficient implementation that could accelerate mainstream adoption of sharded blockchain architectures.
👉 Explore cutting-edge blockchain solutions
FAQs
Q: How does tMPT compare to traditional blockchain synchronization methods?
A: tMPT reduces synchronization overhead by 87% while maintaining complete security guarantees, achieving up to 3.4× higher throughput.
Q: What makes witness nodes different from validator nodes?
A: Witness nodes maintain complete network state and coordinate reorganization, while validator nodes focus on transaction processing within their shards.
Q: How frequently should shard reconfiguration occur?
A: The optimal frequency depends on network conditions, but our protocol demonstrates efficient performance even with frequent reorganizations.
Q: Can tMPT be applied to existing blockchain platforms?
A: Yes, the protocol can be implemented as a layer on top of existing systems, though optimal performance requires architectural adjustments.
Q: What are the security implications of partial state synchronization?
A: Our security proofs demonstrate that partial synchronization maintains network integrity while providing significant performance benefits.
Q: When will blockEmulator be open-sourced?
A: The simulator will be released publicly in the coming months—follow HuangLab for updates.