Research On Key-value Storage Technology Supporting Compression And Encryption In Cloud Computing

Cloud storage services have become the development trend of future storage with its advantages of large storage space and low operating costs with the explosive growth of data volume.With the development of cloud storage,in order to meet the needs of customers,various applications combined with cloud storage continue to improve cloud storage schemes in terms of security,performance,and data access.In order to improve the ability of cloud storage to handle big data workloads,distributed key-value storage has become the backbone of many public cloud services.Modern distributed key-value storage provides superior performance,scalability,and good availability for data-intensive computing and cloud-based applications.Many big data storage systems improve system performance by compressing the data.Data compression allows the server to hold more data in main memory,thereby reducing the number of accesses to persistent storage.Data security has attracted much attention since the advent of cloud computing.Storing data directly on the cloud server will cause the leakage of a large amount of private data.For this reason,encryption technology is usually used to protect the security of data.Therefore,an ideal key-value storage system should incorporate both compression and encryption technologies into the design.However,because compression technology and encryption technology are not easily compatible,most key-value storage systems use only compression or encryption.In addition,the characteristics of the Internet itself also affect the availability of cloud storage.Minimizing communication overhead helps reduce the latency of data transmission.In view of these problems,this thesis researches key-value storage schemes that are compatible with both compression and encryption technologies and support fine-grained access.The main contributions are as follows:First,this thesis proposes a static key-value storage scheme that can support both compression and encryption technologies.Firstly,based on the data division idea of “first horizontally-division,then vertically-division”,a new type of key-value storage structure is designed.This structure uses compression first and then encryption to pack data,and can support column-grained data access.Based on this structure,we combined order revealing encryption technology and order preserving encryption technology to design a secure ordered index in detail,and realized key queries and range queries.The research result has been accepted by the International Conference on Cryptology and Network Security and won the Best Paper Award.Second,In order to enrich the system functions,such as advanced query functions and data update functions,based on the static key-value storage scheme,we combined pseudorandom function technology,searchable symmetric encryption technology,and pessimistic lock ideas to design a general-mode dynamic key-value storage scheme that can support rich queries and dynamic updates.And we combined with the idea of time axial,designed an append-mode dynamic key-value storage scheme for the time series data.Third,we proved the security of the proposed schemes in this thesis,and conducted simulation experiments in a Linux environment.Experimental data shows that the query efficiency of the proposed schemes is acceptable under different query modes,and the query throughput is several times that of the comparison scheme;under the general model,the update throughput is lower,but higher than the comparison scheme;under the append model,the efficiency of dynamic append is very high.To sum up,our schemes are secure and efficient.