Energy optimisation in cloud datacentres with MC-TIDE: Mixed Channel Time-series Dense Encoder for workload forecasting

Haowen Zheng, Yao Lu, Zekun Sun, John Panneerselvam, Xiang Sun and Lu Liu

Applied Energy, 2024, vol. 374, issue C, No S0306261924012868

Abstract: Cloud computing is an integral component of modern IT infrastructure and addressed as an energy consumer due to its increasing utilisation. Cloud workload prediction is regarded as an effective method that can assist cloud service providers with more appropriate resource scheduling, thereby increasing the overall resource utilisation with reduced energy wastage. Herein, accurate prediction models are pivotal to the success of prediction driven energy efficient datacentre management. Despite existing prediction models for cloud datacentres, their accuracy and dependability under limited computational resources still remains a concern due to the resource intensive nature of large prediction models. This study aims to propose different cloud workload statistical methods tailored to the computational resource limitations of various analysts and develop a more comprehensive and accurate cloud computing prediction model based on advanced time-series prediction models to help cloud service providers optimise resource utilisation and reduce energy consumption. Accordingly, we first propose two designs to statistically analyse the cloud workloads: one that is more accurate but consumes more computational resources and the other that simplifies the computation process to require fewer resources while maintaining a certain degree of accuracy. Second, we develop the Mixed Channel Time-series Dense Encoder (MC-TiDE) to efficiently learn information between different time-series, based on the Time-series Dense Encoder (TiDE) model. Experiments conducted on real-world cloud trace logs (including the Alibaba 2018 and Google 2019 datasets) show that our proposed MC-TiDE model outperforms other notable prediction models, demonstrating its prediction accuracy while ensuring efficient training and inference processes.

Keywords: Energy efficiency; Energy saving; Data centre; Cloud computing; Workload prediction; Deep learning (search for similar items in EconPapers)
Date: 2024
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
http://www.sciencedirect.com/science/article/pii/S0306261924012868
Full text for ScienceDirect subscribers only

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:eee:appene:v:374:y:2024:i:c:s0306261924012868

Ordering information: This journal article can be ordered from
http://www.elsevier.com/wps/find/journaldescription.cws_home/405891/bibliographic
http://www.elsevier. ... 405891/bibliographic

DOI: 10.1016/j.apenergy.2024.123903

Access Statistics for this article

Applied Energy is currently edited by J. Yan

More articles in Applied Energy from Elsevier
Bibliographic data for series maintained by Catherine Liu ().