数据因果一致性研究综述

Survey on the causal consistency of data

数据因果一致性是分布式存储中保障数据一致性的重要方案之一,目前的因果一致性方案研究重点包括时钟方法的优化、协议的设计以及操作事务序列的优化等方面。实际上云环境除了时钟漂移、查询放大等情况之外,还存在木马、不可信第三方等不安全因素,以致于破坏因果一致性元数据、用户操作结果的一致性,甚至影响存储环境的可用性。从一致性存储性能提升和安全保障角度出发,结合区块链等共识机制,对比分析了时钟同步方法、数据复制策略、服务端协议的分析设计以及操作事务序列化等研究方向,总结讨论了它们的的原理、优势、局限性以及安全约束方面的不同效用,进而指出未来的发展趋势和后续研究方向,期望对该领域的研究起到参考和帮助作用。

Causal consistency is one of the important projects to ensure data consistency in distributed storage.The current research focuses on causal consistency including optimization of clock method,design of the protocol and the optimization of operation transaction sequence.In the actual cloud environment,in addition to clock skew and query amplification,there are also insecure factors such as Trojans and untrusted third parties,which will destroy the causal consistency metadata stored by users,and the consistency of user’s operating results,even affect the availability of the storage environment.From the perspective of performance improvement and security in distributed storage,the clock synchronization,data replications,analysis and design of server protocol,related research progress of the serialization of operational affairs were introduced combined with consensus mechanisms such as blockchain.At the same time,their principlest,advantages,limitations,and different utilities in terms of security constraints were discussed,and then the future development trends and follow-up research directions were point out at last,which would provide a reference and help for the research in this field.