Enterprises are increasingly adopting multi-cloud infrastructures to increase service availability, scalability, fault tolerance and vendor independence. Nevertheless, the allocation of resources among heterogeneous cloud providers is a highly critical issue because the workloads are variable, the latency is limited, the SLA compliance issues are present, and the privacy concerns are also central to the centralized scheduling models. Conventional centralized reinforcement schedulers in learning demand complete visibility of workload, which can cause overheads in the communication and expose risks of data disclosure. In this paper, we suggest a Federated Reinforcement Learning Scheduler (FRLS) in privacy-sensitive and adaptive optimization of resources on multi-clouds. The proposed architecture has every cloud node separately train a reinforcement learning-based scheduling agent on local observations of workload. Rather than exchanging raw data, nodes exchange model parameters with a federated aggregation server on an irregular basis and weighted averaging is used to create a global scheduling model. Simulations with synthetic heterogeneous workload traces show that FRLS utilizes its resources 15-18 percent more efficiently, SLA violations 8-10 percent fewer and converge quicker than heuristic and centralized RL schedulers. The framework offers scalable, secure and distributed intelligence on next-generation multi-cloud orchestration systems.
Y. Gaidhani, Vudutha Sravanthi, D. L. Narayana et al.· International Conference Com...· 0 citations
The Internet of Things (IoT) ecosystem has expanded and produced massive volumes of real-time streaming data, that must be processed quickly and in an efficient way. The latency problem, bandwidth congestion and bandwidth limitations in the traditional cloud-centric processing architecture are significant challenges, particularly in latency-sensitive systems such as smart healthcare, industrial automation and intelligent transportation systems. While edge computing reduces the amount of data that needs to be sent, edge devices are typically more limited in resources, which means they cannot run complex deep learning models. To address these limitations, this paper proposes a new Edge–Cloud Collaborative Deep Learning (ECC-DL) Framework to orchestrate different aspects of inference and learning between the edge nodes and cloud server. The lightweight convolutional neural networks (CNNs) are deployed at the edge nodes to extract features and make initial inferences, and the heavier end models are deployed at the cloud end for global analysis and regular retraining. To optimize the partitioning of the workload between the resources, an adaptive workload partitioning algorithm is designed that considers latency, bandwidth and computational resources. By experimental evaluation, latency is reduced by 40% and bandwidth consumption is reduced by 30% and prediction accuracy is improved over the centralized approach.
K.M.P.Brundha, Applaneni Lavanya, J.Balaji et al.· 2026 6th International Confe...· 0 citations