Design And Simulated Implementation Of Data Replica Management Mechanism In Cloud Storage For Healthcare Under The IoT Environment

With the rapid development of IoT (Internet of things) technology, A variety of applications based on IoT have become the focus issues in all fields of people’s lives, and healthcare is one of hot areas. The healthcare data under IoT environment has many characteristics, such as mass, heterogeneity, difficulty of management, while the current data replica management mechanism has simple structure and lacks extensibil-ity, for which it can’t provide efficient, reliable and secure data access service for users.The mechanism of data replica management, as a vital part of any storage man-agement system, is especial important to the cloud storage system under the environ-ment of IoT for healthcare. In order to realize the feasibility, reliability and fault toler-ance of the medical data storage, in this thesis a data replica management mechanism based on cloud storage is designed, which includes the data replica placement strategy, selection strategy and consistency maintenance strategy. As for the data replica place-ment strategy, p-center and p-median models are proposed according to the different levels demands of medical data’s access time; the new neighborhood search functions are defined, and the corresponding combination algorithms are designed; due to the dynamism of user access mode and network state, then the replica adjustment strategy is proposed. Taking into account the current state of network and the load of storage sites, a replica selection strategy based on plant growth simulation algorithm is de-signed to accelerate the speed of users accessing medical data. Given the inconsisten-cies caused by the user’s read and write operations to different data replicas, the data replica consistency maintenance strategy is proposed. The access requests can be re-sponded by each data replica and the problem of update conflicts of data replicas is avoided by the timestamp mechanism, which thus improves the system throughput and data access reliability.Finally, a simulation and implementation to the data replica management mecha-nism based on cloud storage proposed in the paper is conducted. The performance as-sessment result shows that, compared with the previous data replica management mechanisms, the mechanism of this paper can not only reduce the time latency and re-source cost, but also achieve the load balance of data replicas and improve the feasibil-ity and reliability of data access at the same time.