Research On Mimic Enhancement Design For Distributed Storage System

With the rapid development of cloud computing and big data technology,massive data continues to emerge.As the cornerstone of cloud storage and big data,distributed storage technology has been widely used in various data fields.As the core element of the storage system,including a large amount of user privacy and sensitive data,how to effectively deal with security issues in the storage field is also of concern to all sectors.Due to its determined storage mode and static storage architecture,the existing distributed storage system is difficult to effectively deal with the end-of-life problem of hardware and software vulnerabilities,which has become the hardest hit area for data loss and leakage.The cyberspace mimetic defense theory proposed by Wu Jiangxing has the characteristics of variable structure and dynamic reconstruction,which provides a new idea for data storage security.To this end,this paper applies the mimetic security defense technology to the distributed storage field,and proposes a new architecture for distributed storage oriented to mimetic enhancement.Improve the security of the storage system by introducing heterogeneity,time-varying,and redundancy.Based on Ceph storage system and distributed fault-tolerant technology,the key technologies of memory system mimicry enhancement are explored and optimized.The main work and achievements are as follows:Research and analysis on the status distributed storage security and the progress of mimicry defense at home and abroad,and discusses the research significance of mimicry defense for distributed storage security.The intrinsic mechanism of the mimicry defense theory,the key technologies of the distributed storage system and the existing problems are deeply studied.Based on the typical Ceph storage architecture,the mimic enhanced security combining the mimic defense and the classical distributed storage theory is proposed.The model design of the mimetic distributed storage system is elaborated.The Ceph architecture is deeply studied in terms of metadata mimetic enhancement,data storage mimetic enhancement,data verification mimetic enhancement,and client mimetic encryption.And the data mimic check throughput was tested.The experimental results show that the mimetic cost is acceptable.Aiming at the storage performance problems caused by the pseudo-hybrid redundancy,a multi-level active queue metadata caching algorithm and a heterogeneous copy combination placement optimization algorithm are proposed.Experiments have shown that optimization of metadata and data copies improves read and write performance in heterogeneous environments.