Dual-Price Hot-Tier Placement and Reliability-Aware Coding for Elastic Cloud Storage Services
Journal of Information, Technology and Policy · 2026
DOI: 10.62836/jitp.v4i1.0002 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of Information, Technology and Policy, 2026. Verfasst von Yizhou Chen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Elastic storage services must allocate hot-tier capacity, move objects across tiers, and maintain durability under changing access patterns. This paper develops DTHC-RS, a dual-price hot-tier controller with reliability-aware erasure coding. The algorithm assigns each object a value density based on predicted access probability, object size, bandwidth benefit, and tail-latency benefit. Two online prices regulate the decision: a capacity price for hot-tier pressure and a reliability price for durability pressure. The theoretical derivation connects the admission rule to a Lagrangian relaxation and connects redundancy choice to the data-loss probability of erasure coding. The case study processes 8,737,496 requests from a public Wikimedia 2019 CDN text trace, observes 2,269,608 anonymous objects, and simulates the first 220,000 requests for hot-tier placement. At 5% hot-tier capacity, DTHC-RS obtains a byte-hit ratio of 20.84% and a P95 TTFB proxy of 0.3959 s. LFU achieves a higher byte-hit ratio in this slice, but DTHC-RS gives a lower tail-latency proxy and a clearer reliability-control interface.
Zitieren
APA
Chen, Y. (2026). Dual-Price Hot-Tier Placement and Reliability-Aware Coding for Elastic Cloud Storage Services. Journal of Information, Technology and Policy. https://doi.org/10.62836/jitp.v4i1.0002
Vancouver
Chen Y. Dual-Price Hot-Tier Placement and Reliability-Aware Coding for Elastic Cloud Storage Services. Journal of Information, Technology and Policy. 2026. doi:10.62836/jitp.v4i1.0002
BibTeX
@article{Chen2026,
author = {Chen, Yizhou},
title = {Dual-Price Hot-Tier Placement and Reliability-Aware Coding for Elastic Cloud Storage Services},
journal = {Journal of Information, Technology and Policy},
year = {2026},
doi = {10.62836/jitp.v4i1.0002}
}
EUVIMED