Abstract: In the distributed storage systems, the metadata stored on the central node, which is prone to a single point of failure and is easily modified by malicious ones. Although the problem can be avoided to some extent by using backup nodes, the synchronization and handover between the nodes are less efficient. At the same time, nodes that store metadata can agree on modifying metadata, which is lacking credibility. For the problems of distributed storage system, a decentralized data storage model is proposed, namely DMB（Decentralized Metadata Blockchain）, which combines with the characteristics of blockchain. The DMB model guarantees metadata integrity by saving metadata in blocks, redundant storage for blockchain, and collaboration validation. The model is divided into two phases which are storage phase and verification phase. In the storage phase, the model sends the user’s signature and locations of data to verification nodes, and uses the signature and data locations to generate metadata block. In the verification phase, the verification nodes check whether the local and global states of the metadata block chain are the same. If they are not the same, the DMB model performs the synchronization of the state, and then the verification nodes retrieve local the metadata blockchain to verify the integrity of metadata. The theoretical and experimental results show that the DMB model can guarantee the retrievability and possession of the metadata, and has better concurrent processing ability, which has little effect on the efficiency of the data stored.

Key words: blockchain, decentration, distributed storage, metadata

摘要： 分布式数据存储过程中的元数据保存在中心节点上，容易造成单点故障和易被恶意修改，安全性较差。虽然，通过引入备份节点可以一定程度上避免该问题，但节点之间的同步和切换效率较低。同时，存储元数据的节点可以达成共识修改元数据，缺乏可信性。针对传统分布式存储中存在的问题，结合区块链的特点，提出一种去中心化的分布式存储模型DMB（Decentralized Metadata Blockchain），通过将元数据保存在区块中、冗余存储区块链、协作验证来保证元数据的完整性。模型分为两个阶段，即元数据存储阶段和元数据验证阶段。在元数据存储阶段，将用户的签名和副本位置数据发送给若干验证节点，生成元数据区块并写入元数据区块链中。在元数据验证阶段，验证节点首先检查本地元数据区块链的状态和全局状态是否相同，如果不相同则进行状态同步。然后，检索本地元数据区块链来验证元数据完整性。理论与实验结果表明，DMB模型可以保证元数据的可追溯性和完整性，有较好的并发处理能力，对数据存储的效率影响较小。

关键词: 区块链, 去中心化, 分布式存储, 元数据