Data deduplication has become a promising technique that significantly reduces space utilization of cloud storage servers and optimizes the network bandwidth. A variety of deduplication techniques have been proposed to enhance the cloud server’s storage efficiency. Most of the deduplication schemes do not address global indexing and user privacy issues. This paper elucidates a novel three-tier secure data deduplication architecture (referred to as SEEDDUP) for data storage and retrieval in the cloud. SEEDDUP provides data deduplication with proper indexing and user privacy. In this work, the optimal chunk size, for a given file size and format, is calculated using Cuckoo Search Algorithm. The chunk existence in the cloud is verified by the hash value generated via SHA3-512. To improve data deduplication speed, Merkle Hash Tree based MapReduce (MHTMR) is used for fast index lookup, which improves the performance at data storage and retrieval. In addition, data are encrypted based on its sensitivity level (low and high) using Niederreiter Public Key Cryptosystem. With this new SEEDDUP architecture, only legitimate users can request data successfully for limited times (k) and spurious data requests cannot be processed or handled in the cloud. Experiment results show that the proposed SEEDDUP can reduce latency, achieved good throughput and deduplication efficiency (chunk size variance and security measures) in data deduplication.
7 February 2022 in Arabian Journal for Science and Engineering
The task scheduling problem in heterogeneous distributed computing systems is a multiobjective optimization problem (MOP). In heterogeneous distributed computing systems (HDCS), there is a possibility of processor and network failures and this affects the applications running on the HDCS. To reduce the impact of failures on an application running on HDCS, scheduling algorithms must be devised which minimize not only the schedule length (makespan) but also the failure probability of the application (reliability). These objectives are conflicting and it is not possible to minimize both objectives at the same time. Thus, it is needed to develop scheduling algorithms which account both for schedule length and the failure probability. Multiobjective Evolutionary Computation algorithms (MOEAs) are well-suited for Multiobjective task scheduling on heterogeneous environment. The two Multi-Objective Evolutionary Algorithms such as Multiobjective Genetic Algorithm (MOGA) and Multiobjective Evolutionary Programming (MOEP) with non-dominated sorting are developed and compared for the various random task graphs and also for a real-time numerical application graph. The metrics for evaluating the convergence and diversity of the obtained non-dominated solutions by the two algorithms are reported. The simulation results confirm that the proposed algorithms can be used for solving the task scheduling at reduced computational times compared to the weighted-sum based biobjective algorithm in the literature.