A popularity-aware reconstruction technique in erasure-coded storage systems

Published: 2020, Last Modified: 21 Jan 2026J. Parallel Distributed Comput. 2020EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: Highlights•The K-prototype and naive-Bayes-based text-analysis algorithms to cluster files during the data archival stage is proposed.•A recommendation system to favor popular data over non-popular ones during the course of data reconstruction.•Real-world traces to evaluate the performance of the POST system is replayed.
Loading