Font Size: a A A

Research On Cloud Data Privacy Preservation Mechanism For Multi-tenancy Applications

Posted on:2013-01-22Degree:DoctorType:Dissertation
Country:ChinaCandidate:K ZhangFull Text:PDF
GTID:1118330374480641Subject:Computer software and theory
Abstract/Summary:PDF Full Text Request
Software as a Service, i.e. SaaS, is one important type of service delivery model in cloud computing. In SaaS, service providers take charge of software maintenance, management and upgrade, while tenants subscribe the software service through web and don't care the implementation detail. Single instance multi-tenancy is the common way adopted by the service providers, by which one instance could serve multiple tenants. The multi-tenancy application has become popular and promising in cloud computing.In multi-tenancy applications, tenants'sensitive data are stored and processed at the platform of un-trustworthy service providers. In order to satisfy the data processing requirements of multi-tenancy applications, data are commonly stored in plain-text form. Data privacy leakage may happen caused by the un-trustworthy service providers from the inside.Data privacy protection for multi-tenancy applicatons faced many challenges, including:(1) In multi-tenancy applications scenario, different tenants have different privacy requirements, and the data privacy preservation should be customized according to the personalized requirements.(2) The multi-tenantcy applications support on-demand customization, and then data privacy preservation should be adaptable to this dynamic scenario and be consistency with the data privacy requirements.(3) The data privacy preservation should combine the data processing performance with data privacy effectively.Considering the data privacy preservation customization support, data privacy preservation consistency, tradeoff between data privacy preservation and processing performance issues, this thesis do research from the following aspects combing the multi-tenancy data sharing architecture, multi-nodes, on-demand and other features of cloud computing and multi-tenancy applications.(1) For sensitive data combination, this thesis proposed data combination privacy preservation mechanism based on information hiding. Combing the balancing technology based on fake tuples, data combination privacy preservation mechanism could protect tenants'data privacy.This thesis proposed a data privacy preservation mechanism based on information hiding, which could protect the sensitive association between data. For privacy leakage caused by data distribution, this thesis proposed the balancing mechanism, which use fake data to adjust the data distribution and prevent the privacy leakage. This mechanism is proved reasonable and is practible after experiments.(2) For single sensitive data in multi-tenancy applications, this thesis proposed non-deterministic data privacy preservation based on supporting vector machine, which could prevent attacks from ciphertext statistics.Considering the issues caused by data encryption and data obfuscation, this thesis proposed non-deterministic data privacy preservation mechanism. This mechanism is based on supporting vector machine and could transform a sensitive data value to many different cipher data value randomly, which could prevent attackers from guessing the original sensitive data value from the cipher text statistics and background knowledge. Analysis and experiments demonstrated the effectiveness and performance of this mechanism.(3) For the privacy leakage caused by on-demand customization of multi-tenancy applications, this thesis proposed the data privacy preservation adjusting mechanism, which could realize the consistency of data privacy requirements of tenants.According to the data privacy preservation requirements, this thesis proposed privacy preservation adjusting mechanism based on data combination privacy. This adjusting mechanism could adjust the data privacy preservation enforcement and minimize the privacy leakage. Analysis and experiments demonstrated the correctness and performance of this data privacy preservation adjusting mechanism.(4) For data storage in cloud, this thesis proposed a data partition and placement mechanism, which could combine the data privacy and performance.Data of tenants could be partitioned and placed in shared storage of multi-nodes in clouds. With the running of multi-tenancy applications, workloads of data nodes changed, which influence the SLA of tenants. This thesis proposed a data partition and placement mechanism, which could combine the privacy and performance.
Keywords/Search Tags:Cloud computing, Multi-tenancy applications, Data privacy, Data combination privacy, Privacy adjusting
PDF Full Text Request
Related items