Jul 2026· 2026 IEEE 9th International Conference on Big Data and Artificial Intelligence (BDAI)· pp. 277-282· 0 citations· 16 references
Abstract
One of the major challenges encountered by cloud computing environments is the ability to handle dynamic workloads at a minimal energy usage and quality of service. This paper discusses a smart resource optimization system, which uses machine learning to forecast resource needs and assign virtual machines to the most efficient tasks. Through a proactive strategy, over-provisioning and under-provisioning that are traps of manual or rule-based scaling are reduced by the system. The method is concentrated on the combination of predictive modelling and automated scaling triggers. The studies apply the CloudSim Plus simulation toolkit as an environment modelling tool and Google Cluster Data traces as the main dataset. In particular, the investigation runs a portion of these traces to recreate heterogeneous real-world task requirements. The outcomes show that energy spending and rejection rates of tasks are significantly lowered as compared to traditional methods of allocation which are not dynamic. The metrics used in the performance are CPU utilization and memory overhead. As in the paper, machine learning models can dramatically improve the operational performance of the large-scale cloud infrastructures by making real-time data-driven decisions, which forms a solid basis behind the next-generation autonomous cloud management systems.
As the cloud computing is developing rapidly, the energy consumption of data centers is growing exponentially and has a significant impact on carbon emissions, which imposes great challenges on the sustainable workload management of cloud platforms. At present, most of the workload allocation strategies in cloud platform are made according to some system performance indicators, such as response time, processing ability, server utilization and so on. However, these indicators cannot reflect the carbon consumption of the power supply of the servers. In this paper, we proposed a framework for thermal-aware and carbon-efficient workload allocation which takes several thermal, energy and environmental-related metrics into account for making decisions on allocating workloads to cloud servers. In order to predict the carbon footprint for running a workload on a server, we trained a carbon emission model which is a Gradient Boosting Regressor and it takes a set of attributes into account when predicting the carbon emissions. For any incoming workload, the thermal-aware allocation engine selects the server with minimum predicted carbon emissions by making use of the model learned in training phase and a set of runtime thermal, energy and environmental metrics. Experiments are designed to evaluate the performance and compare with other allocations. Experimental results show that our framework could effectively select out the most suitable servers for workloads not only from viewpoint of carbon usage but also from the thermal perspectives, and at the same time, it also reduces the amount of cooling required and carbon emissions while keeps several performance metrics at comparable levels. Our work shows that by taking into account the thermal metrics and the current carbon intensity information, we could make sustainable scheduling decisions in cloud platforms.
Chandan Hegde, Adarsh Bilimisi, Pruthvik J· International Journal of Lat...· 0 citations
The rapid evolution of machine learning (ML) models and the surge in data volumes necessitate scalable and efficient deployment strategies. Cloud-based distributed systems offer on-demand scalability and resource flexibility, making them ideal for real-time ML model deployment and scaling. This paper explores optimization techniques for cloud-based distributed systems to enhance the deployment and scaling of ML models in real-time applications. We examine the integration of distributed systems and ML within cloud environments, focusing on scalable training and inference mechanisms. Key considerations such as task partitioning, communication overhead, fault tolerance, and resource optimization are discussed. Furthermore, we review auto-scaling techniques, highlighting advancements and challenges in dynamically adjusting resources to meet fluctuating demands. The paper also delves into the application of machine learning for cloud resource provisioning, emphasizing dynamic allocation based on real-time usage patterns. By synthesizing current research and practices, this study provides insights into effectively leveraging cloud-based distributed systems for real-time ML model deployment and scaling.
Emma Roberts, William Hughes· International Journal of Art...· 0 citations
For companies with a multi-cloud strategy, cost management has turned into a real challenge. It is not just a matter of tracking costs; existing tools provide a retrospective view of costs, not a prospective one, which makes planning difficult. In this paper, the FinOps-based Intelligent Serverless Orchestrator (FISO) is described as a real-time predictive system based on machine learning and streaming data processing, which helps mitigate cost overruns before the financial run rate is impacted. FISO achieves a forecast accuracy of 93.4% (MAPE: 8.3%) and detects cost anomalies within 4 minutes, compared to 24–48 hours with conventional tools. It combines forecasting (using the Facebook Prophet algorithm) and cost anomaly detection (Isolation Forest) with a natural language interface, enabling users to query costs conversationally. FISO was evaluated on real production data spanning 1–2 months across AWS, Azure, and Google Cloud, with the system ingesting data every 2 minutes. API calls remained under 200 ms while processing thousands of metrics per minute. Most importantly, organizations using FISO reduced budget surprises by 64% and improved forecast accuracy from 69.8% to 93.4%. The main contributions are: (1) a practical framework for normalizing cost data across disparate cloud providers; (2) a machine learning pipeline that self-adjusts to incorporate shifting usage trends; (3) a natural language dialogue interface for financial cost queries; and (4) empirical evidence of production-level performance confirming reliability in operational settings.
Sameer Krishn Sistla, S. Tilak, B. V· International Conference Com...· 0 citations
Reinforcement learning-based adaptive resource management framework is proposed that enables cloud systems to autonomously learn optimal resource allocation policies through continuous interaction with the environment and significantly outperforms static and reactive baseline strategies in terms of resource utilization efficiency and response time stability.
Rajesh Sharma, Priya Natarajan· International Journal of Mac...· 0 citations
Adaptive Machine Learning Cloud Services are capable of adjusting their structure or behavior. They may, for example, add or remove components based on capacity demands or improve the quality of the service with more accurate algorithms. Furthermore, it is possible to adapt the structure of a Machine Learning architecture so that the inputs and outputs relate correctly for a particular task, such as classifying objects that have been detected in an image. Although there is significant literature describing Machine Learning as a Service, there is still little work that considers the specific needs of self-adaptive Machine Learning systems. In a cloud context, self-adaptation can take place in a service-oriented architecture and several factors can trigger this self-adaptation, including system or user needs. A guiding principle for resource management is elasticity, the ability to adapt the level of computing resources allocation to the current demand. Therefore, it should also be possible to manage the internal resources of a self-adaptive Machine Learning Service. A cornerstone concern is that the adjustment of the service requires monitoring. To avoid the overhead of monitoring particularly when resources are idle, a different approach is employed here, focusing on the reference behavior of a process. Instead of the intrinsic properties measured in several other monitoring papers, external indicators of quality are coupled to a feedback loop that triggers adjustment of the internal behavior.
Unknown authors· American International Journ...· 0 citations
Data centers in the cloud use a lot of energy and produce considerable carbon emissions because of the growing requirements for computational and AI-heavy processes. Currently, available cloud schedulers pay attention to performance and resource usage optimization without much consideration of carbon footprint efficiency, security compliance, and region-based regulations for cloud data centers. In this research paper, we suggest creating a Secure Carbon-Aware Scheduler that includes machine learning-based carbon intensity prediction, prioritization of the workload, and compliance-based scheduling for distributed cloud environments. Our solution includes using a hybrid approach to carbon monitoring by integrating live carbon intensity data from one pilot region with predicted data based on machine learning models for distributed regions. We trained three ML algorithms such as Random Forest, Gradient Boosting, and Linear Regression for three years of data on carbon intensity, and the best models were chosen according to R2 score and MAE metric.The scheduler performs an assessment of potential regions through a multi-criteria objective scoring function that is based on carbon footprint, security compliance, workload prioritization, and system performance. The architecture employs security enforcement methods that support policies such as GDPR and HIPAA, as well as asymmetric cryptography-based security features for sensitive workloads. Experiments have been conducted by simulating cloud workloads from fourteen different regions and showed promising results in terms of workload prioritization and better carbon-aware scheduling when compared to conventional scheduling algorithms.
Shreenidhi Anant Bhat, Premalatha H. M.· 2026 4th International Confe...· 0 citations