Data repair optimization method for geo-distributed fault-tolerant storage systems based on whale optimization algorithm

Erasure codes have been gradually adopted in today’s geo-distributed cloud data centers to ensure cross-regional data reliability while minimizing storage overhead. However, unexpected node failures trigger costly cross-data-center repair operations, resulting in substantial network traffic, extended recovery times, and uneven system loads. Current repair strategies fail to effectively balance three critical factors: repair latency, traffic overhead, and load distribution. To address these challenges, we propose a comprehensive data repair optimization method based on the Whale Optimization Algorithm (WOA) for geo-distributed erasure-coded storage. Our approach begins by formulating a multi-objective optimization model that jointly considers repair latency, traffic overhead, and load balancing. We then develop a WOA-based topology construction algorithm to identify near-optimal repair solutions within the Pareto frontier. Furthermore, we integrate a Dijkstra-based loop elimination technique to ensure efficient, loop-free repair paths. Extensive simulations demonstrate that our method achieves a reduction of up to 41.11% in average repair latency and an 86.43% decrease in link utilization standard deviation compared to baseline algorithms, while maintaining minimal repair traffic overhead. These results prove its superior effectiveness in optimizing repair performance and load balancing for geo-distributed storage systems.