一种拟态存储元数据随机性问题解决方法

拟态存储作为网络空间拟态防御技术在分布式存储领域的工程实现，能够有效提高存储系统的安全性。由于元数据节点中存在随机性的算法和逻辑，使得执行体状态可能不一致，从而导致整个拟态存储系统无法正常运转。为解决该问题，提出一种元数据再同步方法。在系统中引入状态监视模块和映射同步机制，状态监视模块及时检测执行体状态不一致的情况并进行反馈，映射同步机制建立客户端指令和裁决器输出结果之间的映射关系，使得同步过程在不影响拟态存储系统正常工作的同时符合分布式存储系统最终的一致性要求。在拟态存储工程样机上进行功能和性能测试，结果表明，该方法能够以较小的性能开销来有效解决元数据执行体状态不一致的问题，提升拟态存储系统的稳定性，重复执行再同步机制可使同步成功率达到100%。

A Method for Solving the Metadata Randomness Problem of Mimic Storage

Mimic storage is an engineering implementation of cyberspace mimic defense technology in the field of distributed storage, and can effectively improve the security of the storage system.Due to the random algorithm and logic in the metadata node, there is a probability that states of the executor are inconsistent, affecting the operation of the whole mimic storage system.To address the problem, a metadata resynchronization method is proposed.We introduce a state monitoring module and a mapping synchronization mechanism into the system.The sate monitoring module detects and reports the inconsistency of executor states.The mapping synchronization mechanism builds the mapping relationship between the client instructions and the arbitrator output, which allows the synchronization process meet the eventual consistency requirement of distributed storage system without interfering with the operation of mimic storage system.Functional tests and performance tests are carried out on a mimic storage engineering prototype.Results show that this method can effectively solve the inconsistency of metadata executors at a small performance loss.It improves the stability of mimic storage system, and increases the synchronization success rate to100% by repeating the resynchronization mechanism.