CFWS: DRL-Based Framework for Energy Cost and Carbon Footprint Optimization in Cloud Data Centers
Open accessThe DOI matches its Crossref record; 33 of the 33 references it cites were found in a registry or an index.
Abstract
The rapid growth and widespread adoption of cloud computing have led to significant electricity costs and environmental impacts. Traditional approaches that rely on static utilization thresholds are ineffective in dynamic cloud environments, and simply consolidating virtual machines (VMs) to minimize energy costs does not necessarily result in the lowest carbon footprints. In this paper, a deep reinforcement learning (DRL) based framework called CFWS is proposed to enhance the energy efficiency of renewable energy sources (RES) supplied data centers (DCs). CFWS incorporates an adaptive thresholds adjustment method TCN-MAD by evaluating the predicted probability of a physical machine (PM) being overloaded to prevent unnecessary VM migrations and mitigate service level agreement (SLA) violations due to imbalanced workload distribution. Additionally, CFWS introduces a novel action space in the DRL algorithm by representing VM migrations among geo-distributed cloud data centers as flattened indices to accelerate its execution efficiency. Simulation results demonstrate that CFWS can achieve a superior optimization of energy costs and carbon footprints, saving 5.67% to 13.22% brown energy with maximized RES utilization. Furthermore, CFWS reduces VM migrations by up to 86.53% and maintains the lowest SLA violations within suboptimal execution time in comparison to the state-of-art algorithms.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
An Empirical Evaluation of Generic Convolutional and Recurrent Networks for Sequence Modeling
Bai, S., Kolter, J. Z., & Koltun, V. (2018) · arXiv · doi:10.48550/arxiv.1803.01271 · Cited by 4,655 · cited 1×
“It is also expected to consider the impact of cooling and network transmission on energy consumption to prevent service quality degradation due to insufficient RES supply.”
Beloglazov, A., Abawajy, J., & Buyya, R. (2012) · Future Generation Computer Systems · doi:10.1016/j.future.2011.04.017 · Cited by 2,761 · cited 1×
“[26] shows that DCs can obtain excellent performance when the lower threshold is 40%, regarded as μ Lk = 40% .”
Beloglazov, A., & Buyya, R. (2012) · Concurrency and Computation: Practice and Experience · doi:10.1002/cpe.1867 · Cited by 1,864 · cited 2×
“The upper threshold of the PMs allocated in the k th DC can be defined by where the value of s is 2.5 through experimental results mentioned in [25].”
InfluentialSmartly Handling Renewable Energy Instability in Supporting A Cloud Datacenter
Gao, J., Wang, H., & Shen, H. (2020) · 2020 IEEE International Parallel and Distributed Processing Symposium (IPDPS) · doi:10.1109/ipdps47924.2020.00084 · Cited by 272 · cited 1×
“Although RES is assumed togenerate0carbonemission[22],thispaperfollowsthesettings in[23],[24],consideringamorerealisticscenariowherethe CFRvalueofwindistreatedasaconstant.”
Machine learning (ML)-centric resource management in cloud computing: A review and future directions
Khan, T., Tian, W., Zhou, G., Ilager, S., Gong, M., & Buyya, R. (2022) · Journal of Network and Computer Applications · doi:10.1016/j.jnca.2022.103405 · Cited by 179
Yadav, R., Zhang, W., Kaiwartya, O., Singh, P. R., Elgendy, I. A., & Tian, Y.-C. (2018) · IEEE Access · doi:10.1109/access.2018.2872750 · Cited by 177 · cited 1×
“Toward this goal, the over-utilization resource can be migrated to a limited number of active PMs and the remaining PMs can be switched to the inactive mode for saving energy [4].”
Khosravi, A., Andrew, L. L. H., & Buyya, R. (2017) · IEEE Transactions on Sustainable Computing · doi:10.1109/tsusc.2017.2709980 · Cited by 141 · cited 1×
“The representative research [20] calculates the PUE as (7) Furthermore, the P IT k ( t ) in (6) can be calculated by summing up all servers’ power consumption in the k th DC, which can be formalized as Considering that constructing an accurate PM energy model is quite complicated, the SPECpower…”
Xu, C., Wang, K., Li, P., Xia, R., Guo, S., & Guo, M. (2018) · IEEE Transactions on Network Science and Engineering · doi:10.1109/tnse.2018.2813333 · Cited by 124
Shaw, R., Howley, E., & Barrett, E. (2022) · Information Systems · doi:10.1016/j.is.2021.101722 · Cited by 112
Managing renewable energy and carbon footprint in multi-cloud computing environments
Xu, M., & Buyya, R. (2019) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2019.09.015 · Cited by 108 · cited 1×
“Xu et al. [13] designed an innovative workload management strategy that addresses the challenge of carbon emissions by prioritizing cloud DCs with sufficient RES or lower CFRs in multi-cloud environments.”
Utilizing power consumption and SLA violations using dynamic VM consolidation in cloud data centers
Arshad, U., Aleem, M., Srivastava, G., & Lin, J. C. (2022) · Renewable and Sustainable Energy Reviews · doi:10.1016/j.rser.2022.112782 · Cited by 103 · cited 2×
“Arshad et al. [6] presented a dual-threshold approach to classify hosts into three primary categories by interquartile range, which can effectively capture and analyze different levels of host utilization for improved energy-related management.”
Adaptive DRL-based Virtual Machine Consolidation in Energy-Efficient Cloud Data Center
Zeng, J., Ding, D., Kang, X. K., Xie, H., & Yin, Q. (2022) · IEEE Transactions on Parallel and Distributed Systems · doi:10.1109/tpds.2022.3147851 · Cited by 100 · cited 2×
“(cid:2) ADVMC: Thismethodisderivedfrom[16],whichdevelops a DRL-based VM placement strategy and adopts LSTM to predict the system status in a traditionally powered cloud data center.”
Green Cloud Framework for Improving Carbon Efficiency of Clouds
Garg, S. K., Yeo, C. S., & Buyya, R. (2011) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-23400-2_45 · Cited by 99 · cited 1×
“The US Department of Energy Electricity Emission Factors [31] provides the carbon footprint rates of the four data centers, which are presented in Table V.”
An adaptive heuristic for managing energy consumption and overloaded hosts in a cloud data center
Yadav, R., Zhang, W., Li, K., Liu, C., Shafiq, M., & Karn, N. K. (2018) · Wireless Networks · doi:10.1007/s11276-018-1874-1 · Cited by 97
Self-citationManaging overloaded hosts for energy-efficiency in cloud data centers
Yadav, R., Zhang, W., Li, K., Liu, C., & Laghari, A. A. (2021) · Cluster Computing · doi:10.1007/s10586-020-03182-3 · Cited by 84 · cited 1×
“Yadav et al. [9] introduced the stochastic gradient descent method for effectively detecting overloaded hosts, while also designing an energy-aware VM selection policy based on predicted minimum utilization.”
Self-citationMapetu, J. P. B., Kong, L., & Chen, Z. (2020) · The Journal of Supercomputing · doi:10.1007/s11227-020-03494-6 · Cited by 72 · cited 1×
“The table presents nine combinations using different overloaded detection algorithms, including TCN-MAD, LSTM-MAD, MAD, IQR, and THR (a static threshold set to 0.8 [33]).”
InfluentialTwo-Phase Industrial Manufacturing Service Management for Energy Efficiency of Data Centers
Zhang, W., Yadav, R., Tian, Y., Tyagi, S. K. S., Elgendy, I. A., & Kaiwartya, O. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3153508 · Cited by 71 · cited 1×
“Zhang et al. [8] proposed an enhanced adaptive threshold classification principle using the least median square regression technique, which enables resource migration among four distinct groups for optimal SLA compliance and energy utilization.”
Fine-Grained Energy Consumption Model of Servers Based on Task Characteristics in Cloud Data Center
Zhou, Z., Abawajy, J., Li, F., Hu, Z., Chowdhury, M. U., Alelaiwi, A. A., & Li, K. (2017) · IEEE Access · doi:10.1109/access.2017.2732458 · Cited by 61 · cited 1×
“…up all servers’ power consumption in the k th DC, which can be formalized as Considering that constructing an accurate PM energy model is quite complicated, the SPECpower benchmark [21] is adopted to evaluate P ITjk , which is decided by the j th server’s CPU utilization, as shown in Table II.”
Self-citationRawas, S. (2021) · Multimedia Tools and Applications · doi:10.1007/s11042-021-10616-6 · Cited by 40
Wei, P., Zeng, Y., Yan, B., Zhou, J., & Nikougoftar, E. (2023) · Journal of King Saud University - Computer and Information Sciences · doi:10.1016/j.jksuci.2023.04.002 · Cited by 39
Self-citationPower and thermal-aware virtual machine scheduling optimization in cloud data center
Chen, R., Liu, B., Lin, W., Lin, J., Cheng, H., & Li, K. (2023) · Future Generation Computer Systems · doi:10.1016/j.future.2023.03.049 · Cited by 33
Self-citationProactive auto-scaling for cloud environments using temporal convolutional neural networks
Golshani, E., & Ashtiani, M. (2021) · Journal of Parallel and Distributed Computing · doi:10.1016/j.jpdc.2021.04.006 · Cited by 32 · cited 1×
“Since TCN has been shown to have better accuracy while predicting the workload variation than LSTM [34], [35], TCN-based methods reduce subsequent migrations to rebalance the workload (eg.”
Minimizing Energy Cost for Green Data Center by Exploring Heterogeneous Energy Resource
Hu, X., Li, P., & Sun, Y. (2021) · Journal of Modern Power Systems and Clean Energy · doi:10.35833/mpce.2019.000052 · Cited by 31 · cited 1×
“Hu et al. [15] presented a two-stage method to address the energy variances caused by geo-distributed RES generators, which quantifies the greenness of each energy source using the average carbon emission rate and establishes a distribution power model to minimize total energy costs.”
Patchell, J., & Hayter, R. (2020) · Journal of Cleaner Production · doi:10.1016/j.jclepro.2020.125501 · Cited by 25 · cited 1×
“IT giants such as Apple and Facebook have both achieved carbon neutrality through their solar-supplied data centers [5].”
Towards Energy Efficiency and Power Trading Exploiting Renewable Energy in Cloud Data Centers
Aslam, S., Aslam, S., Herodotou, H., Mohsin, S. M., & Aurangzeb, K. (2020) · International Conference on Advances in the Emerging Computing Technologies (AECT) · doi:10.1109/aect47998.2020.9194169 · Cited by 20
Multi-Agent Reinforcement Learning based Distributed Renewable Energy Matching for Datacenters
Wang, H., Shen, H., Gao, J., Zheng, K., & Li, X. (2021) · Proceedings of the International Conference on Parallel Processing · doi:10.1145/3472456.3473514 · Cited by 20
Lu, X., Jiang, D., He, G., & Yu, H. (2018) · Sustainable Computing Informatics and Systems · doi:10.1016/j.suscom.2018.07.004 · Cited by 12
LECC: Location, energy, carbon and cost-aware VM placement model in geo-distributed DCs
Rawas, S., Zekri, A., & El-Zaart, A. (2022) · Sustainable Computing: Informatics and Systems · doi:10.1016/j.suscom.2021.100649 · Cited by 12 · cited 2×
“Rawas et al. [11] proposed a location-aware VM consolidation approach (LECC) for geo-distributed cloud DCs, which evaluates several overloaded detection methods in advance and then selected the minimum carbon and cost data center for migrating VMs.”
Planning for green cloud data centers using sustainable energy
Gu, C., Li, Z., Liu, C., & Huang, H. (2016) · IEEE Symposium on Computers and Communications (ISCC) · doi:10.1109/iscc.2016.7543835 · Cited by 10 · cited 1×
“In contrast to previous studies that consider both solar and wind energy [28], this paper assumes that NE-3000 wind turbines [30] are installed at each data center to fulfill the electric necessities of cloud DCs. Table IV provides an overview of the wind turbine parameters, and the generated wind…”
Online VM Consolidation in Cloud Environments
Alsadie, D., Tari, Z., & Alzahrani, E. J. (2019) · IEEE ... International Conference on Cloud Computing · doi:10.1109/cloud.2019.00033 · Cited by 9 · cited 2×
“To be specific, each data center has a varying number of heterogeneous PMs, with half of them being HP ProLiant ML G4s and the other part HP ProLiant ML110 G5s [28].”
Renugadevi, T., & Geetha, K. (2021) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-189887 · Cited by 9 · cited 1×
“Renugadevi et al. [14] proposed an optimization function that considers electricity and carbon costs under task deadline constraints, which incorporates the concept of application brownout and batch task delays to maximize RES utilization.”
Nadalizadeh, Z., & Momtazpour, M. (2021) · The Journal of Supercomputing · doi:10.1007/s11227-021-03891-5 · Cited by 9 · cited 2×
“(cid:2) Greenpacker: This method is derived from [12], which focuses on RES-powered data centers.”
Green-Aware Virtual Machine Migration Strategy in Sustainable Cloud Computing Environments
Wang, X., Zhang, G., Yang, M., & Zhang, L. (2017) · InTech eBooks · doi:10.5772/67350 · Cited by 4
Research works printed but not linked: 1
- Than, 2015, Estimated social cost of climate change not accurate, 12, 1391
Cited by
Kahil, H., Sharma, S., Välisuo, P. O., & Elmusrati, M. S. (2025) · Applied Energy · doi:10.1016/j.apenergy.2025.125734 · Cited by 73
Power Management Optimization for Data Centers: A Power Supply Perspective
Huang, H., Lin, W., Lin, J., & Li, K. (2025) · IEEE Transactions on Sustainable Computing · doi:10.1109/tsusc.2025.3542779 · Cited by 34
“Whereas in research based on RL, CFWS [122] and DeepScale [123] were proposed to solve the VM scheduling and container expansion problems in virtualization technology.”
Self-citationAlex, M., Ojo, S. O., & Awuor, F. M. (2025) · Computers · doi:10.3390/computers14070280 · Cited by 22
Adaptive Multi-Objective Virtual Machine Consolidation for Energy-Efficient Cloud Data Centers
Goyal, S., & Awasthi, L. K. (2025) · Journal of Grid Computing · doi:10.1007/s10723-025-09808-3 · Cited by 15
“Although some studies [23, 25– 28] have examined the relationship between CPU utilization and VM performance degradation, however, does not present a complex relation between VM consolidation and VM performance degradation.”
Liu, Q., Pan, Z., Guo, Y., Chen, Y., Štrbac, G., & Qiu, D. (2026) · Applied Energy · doi:10.1016/j.apenergy.2026.127561 · Cited by 7
Metadata
- Authors
- Zhao, D., Zhou, J.-T., & Li, K.
- Affiliations
- SUNY New Paltz (US)
- College of Computer Science, Inner Mongolia University, Hohhot, China
- State University of New York (US)
- Venue
- IEEE Transactions on Sustainable Computing
- Published
- 2025-01
- Volume
- 10
- Issue
- 1
- Pages
- 95-107
- ISSN
- 2377-3782 (ISSN-L)
- 2377-3790
- OpenAlex
- W4395017660
- Cited by
- 32
- References
- 33
- Citations by year
- 2026: 21 · 2025: 11
- FWCI
- 11.98
- Type
- journal article
- Language
- en
- Open access
- hybrid
- doi.org
- Locations
- IEEE Transactions on Sustainable Computing (publishedVersion, cc-by)
- Funders
- National Natural Science Foundation of China (62162046)
- Natural Science Foundation of Inner Mongolia Autonomous Region (2019ZD15)
- Research and Application of Key Technology of Big Data for Discipline Inspection and Supervision (2019GG372)
- Science and Technology Plan Special Project of Hohhot
- Topics
- Physical Sciences › Computer Science › Information Systems › Cloud Computing and Resource Management
- Physical Sciences › Computer Science › Hardware and Architecture › Energy Efficiency in Computing
- Physical Sciences › Engineering › Electrical and Electronic Engineering › Green IT and Sustainability
- Keywords
- cloud data centers, energy efficiency, carbon footprint, deep reinforcement learning, virtual machine migration, renewable energy, service level agreement violation, workload distribution, geo-distributed cloud, energy cost optimization
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.