Research Of Sensor-cloud Data Retrieval And Privacy Protection Based On Fog Computing

The development of the internet plus,the internet of things and big data pushes the wireless sensor networks(WSNs)in cloud environment becoming the new kinetic energy of technological innovation and the new engine of system architecture expansion.To meet the high-quality requirements of users for resources and configuration,the sensor-cloud environment has become one of the preferred directions for system and data migration.Sensor-cloud systems can offset the defect of sensor nodes in their resource and performance,providing more powerful computing,storage,processing and analysis capabilities,and avoiding the data task overload of the underlying network.Compared with the traditional WSNs,the mobile nodes based on cloud computing and the state change of wireless networks are highly dynamic.Moreover,wide openness is an inherent feature of the cloud computing environment.These factors determine that the massive data in the sensor-cloud network is facing a very high risk of privacy disclosure,and also brings some obstacles to the construction of data security and privacy protection mechanism.Through the comprehensive analysis of sensor-cloud data application,the security threats existing in data storage and retrieval by users using the cloud platform.It becomes the main problem to be solved in privacy protection mechanism:(1)There is a privacy threat to end-user data transmission to the cloud.Because the data is completely outsourced and stored with the cloud server,the user as the data owner will lose control and management of the stored data.Meanwhile,users need to query and obtain data from the cloud when necessary.This process is a high-risk link for privacy disclosure.In the present research,the privacy protection of the data flow process has not received much attention.(2)There is a lack of information protection mechanisms for data prefetching in the cloud.In the data information publishing scenario,besides the risk of data being leaked and tampered,the access policy also improves the possibility of end-user information exposure to some extent.(3)There is a lack of reliable management of data sharing in sensor-cloud systems.Cloud storage applications include several personal information,enterprise sensitive information,and related state data.End-user need a more reliable data retrieval and sharing scheme to avoid the serious decision-making defects caused by malicious sharing.Fog computing technology proposes to set up high-performance mobile nodes between WSNs and cloud platforms to assist in the expansion of the cloud.It brings processes such as data processing,computing,and storage in the vicinity of end-user together.Rather than fully delivering to the cloud,the fog platform is built in the form of a distributed architecture and an advantaged feature closer to the edge of the network.Compared to a single cloud environment,fog computing provides more direct management and control capabilities to compensate for the cloud platform's shortcomings in service distance.Based on this,this paper studies the introduction of fog servers in sensor-cloud systems to manage and enhance privacy protection performance.The main researches are as follows:(1)A hierarchical data partitioning strategy of "local-fog-cloud" is offered,which can realize semantic security based on information quantity.Data blocks partitioned by this policy are stored in different layers of servers.There is not any piece of data can be used independently for raw data recovery which ensures data privacy.Correspondingly,the retrieval index between the local user and fog servers is built implicitly to avoid the privacy disclosure caused by the exposure of the index content.(2)Using the characteristics of fog nodes in computing and storage,an active bidirectional prefetching based on fog-agent is designed.It can manage the local use directly,and construct a deep reinforcement learning model for highly dynamic and complex user behavior: Fog servers active individually or jointly to prefetching content from the cloud,and extract the user's playback behavior to determine prefetching content to local users.(3)A private data sharing mechanism based on fog computing is proposed to achieve combination sharing.Through virtualized shared data sets established at the fog end,it can improve the availability and privacy of data management in sensor-cloud systems.On the one hand,fog nodes are involved in the implementation of adaptive privacy policy generation,such as open operation permission;on the other hand,the sharing model can take into account the privacy requirements of the data provider and the data user.The privacy protection mechanism in this paper uses fog computing to design a three-layer structure.It forms collaborative services with cloud servers to improve data privacy and cloud service quality.Based on this framework,the security of hierarchical data retrieval and sharing is realized,This paper provides detailed performance analysis and extensive experiments for different data transfer scenario settings to demonstrate the effectiveness of the framework for data privacy protection.