《计算机应用》唯一官方网站

• •    下一篇

基于可信执行环境的自适应在线区块链分片方法

王飞1,王恒笛1,朱孔林1,董进2,张琳1   

  1. 1. 北京邮电大学
    2. 北京未来区块链与隐私计算高精尖创新中心
  • 收稿日期:2024-12-31 修回日期:2025-02-06 发布日期:2025-03-13 出版日期:2025-03-13
  • 通讯作者: 王飞
  • 基金资助:
    国家重点研发计划“区块链”重点专项

Adaptive online blockchain sharding method based on trusted execution environments

  • Received:2024-12-31 Revised:2025-02-06 Online:2025-03-13 Published:2025-03-13

摘要: 摘 要: 分片是一种被广泛采用的增强区块链交易处理能力的解决方案。然而,现有的跨分片交易协议通常需要多轮分片到分片的通信,严重影响了系统性能。本文提出了自适应在线分片方法,这是一种创新的区块链框架,它利用概率延迟机制和可信执行环境来优化复杂的跨分片交易的执行。首先,自适应在线分片战略性地延迟交易分配到分片,使相关交易能够分组在一起,从而减少跨片交易的数量并最小化通信开销。然后,自适应 在线分片利用可信执行环境安全有效地执行链下的跨片交易,从而消除了多轮通信的需要。本文的分析表明,自适应在线分片实现了次线性动态后悔,适应不断变化的交易模式的能力。此外,自适应在线分片专为动态在线环境而设计,不断适应实时交易数据和反馈。并在真实世界的区块链设置中评估了它的性能,证明了自适应在线分片显著优于最先进的分片方法,提高了30%吞吐率。

关键词: 区块链, 分片, 可信执行环境, 在线优化, 跨片交易

Abstract: Abstract: Sharding has emerged as a widely adopted solution to enhance the transaction processing capabilities of blockchain systems. However, existing cross-shard transaction protocols often necessitate multiple rounds of communication between shards, which significantly hampers system performance. This paper presents an Adaptive Online Sharding Method, an innovative blockchain framework that leverages probabilistic delay mechanisms and Trusted Execution Environments (TEEs) to optimize the execution of complex cross-shard transactions. The proposed method strategically delays the allocation of transactions to shards, enabling the grouping of related transactions and thereby reducing the occurrence of cross-shard transactions while minimizing communication overhead. Additionally, it utilizes TEEs for secure and efficient off-chain execution of cross-shard transactions, effectively eliminating the need for multiple communication rounds. Our analysis shows that the Adaptive Online Sharding achieves sub-linear dynamic regret, demonstrating its adaptability to evolving transaction patterns. Furthermore, the framework is designed for dynamic online environments, continuously adjusting to real-time transaction data and feedback. Performance evaluations in real-world blockchain settings indicate that the Adaptive Online Sharding significantly outperforms state-of-the-art sharding methods, improving the TPS by 30%.

Key words: blockchain, sharding, trusted execution environments (TEE), online optimization, cross-shard transactions

中图分类号: