ENHANCING EDGE DATA DEDUPLICATION WITH ROBUST OPTIMIZATION AMIDST UNCERTAINTIES

Abstract

The rise of mobile edge computing (MEC) within distributed systems has brought greater focus to managing data at the network edge. A significant challenge exists due to the limited storage capacity of edge servers contrasted with the everincreasing demand for data storage, making cost reduction a key priority. Although edge data deduplication has been widely explored as a method for reducing data redundancy, current approaches face various obstacles in MEC settings. These difficulties arise from differences between edge servers and traditional cloud data centers, as well as unpredictable factors like user mobility, which undermine the reliability of deduplication strategies. To address these issues, this paper introduces a robust optimizationbased framework for edge data deduplication. By incorporating uncertainties such as fluctuating data demands and potential edge server failures, we develop two solution algorithms: uEDDE-C, a twostage method leveraging column-andconstraint generation, and uEDDE-A, an approximation technique designed to reduce the computational complexity of uEDDE-C. Our approach enables effective data deduplication in dynamic edge environments while maintaining resilience under various uncertain conditions. We substantiate the efficacy and stability of both algorithms through rigorous theoretical analysis and comprehensive experiments. The results confirm that our method substantially lowers data storage expenses and decreases data retrieval latency, ensuring dependable performance in practical MEC deployments.