Research On Key Technologies Of Network-Coding-Based Mimic Data Storage

With the developement of informationization in China,information technology has been widely used in different areas.The emergence of the mobile Internet,social network,and electronic business greatly extends the application area of the Internet.As rapid inflation and accumulation of various type of data in different fields,computing and storage mode is faced with enormous challenges.Cloud storage technology provides a good solution to store and process big data.Nowadays,more and more users choose to outsource their own data on cloud storage platform,cloud storage system has become an important information infrastructure in modern society.However,the existing cloud storage technologies are still facing a series of security problems,these problems seriously restrict the further development and application of cloud storage.Existing cloud storage system become a major disaster area of network attack and data leakage due to the certainty of storage mode.Traditional security mechanisms of cloud storage are based on static storage architecture and use passive defensive strategy,it's difficult to defense against the increasingly complicated and diverse network attacks.Therefore,this paper applies mimic security defense technology to the field of cloud storage,by introducing redundancy,time-varying and randomness to improve the security and reliability of the storage system.We do some exploration and research on the key technologies of mimic data storage based on network-coding technology in this paper.The main work is as follows:1.Aiming to solve security threats in the cloud storage system due to static storage architecture and storage mode,this paper proposes a mimic storage scheme based on regenerated code.The scheme uses network coding technology to store the data in the data nodes and uses mimic transformation mechanism based on regenerated code to change storage status of outsourced data according to the random and time-varying factors dynamically,which can always guarantee data integrity and data availability.Mimic transformation mechanism is a random,time-varying and dynamic scheme,which increases the uncertainty of storage system.It can block and interference attack chain,increase the difficulty and cost of the attack operation greatly,and improve the security and reliability of the system.2.In order to address the problem that the communication cost of update operation on the network-coding-based storage system is too large,a differential data update scheme is proposed.The scheme only encodes and updates the change parts in the file,which can significantly reduce the bandwidth overhead of the update process.According to the redundant characteristic of the encoded data,we propose a joint compression algorithm by combining Run Length coding and Huffman coding,which has better compression effect,and can be used to further reduce the communication overhead of the update process.Compared with the existing scheme,the bandwidth overhead and computational overhead of this scheme are significantly reduced,and the data node is not required to have the ability of coding and decoding,this scheme has better adaptability and compatibility.3.A network-coding-based mimic storage verification system is designed and implemented,which implements the mimic storage scheme and the differential data updating scheme mentioned above using network-coding technology.The system adopts the layered and loose-coupled structure to support functional extension and dynamic reconstruction,and has good scalability and compatibility.In the real network and storage environment,we verify the functions of the mimic storage scheme and the differential data update scheme,and do some related performance tests,the results show that the mimic storage scheme has good feasibility and adaptability,and the differential data update scheme can effectively reduce the bandwidth overhead of data update process.