Recent Advances in Joint Resource Allocation, Security, and Efficient Task Scheduling in Cloud Computing Using Hybrid Pyramidal Convolution Split-Attention Networks: A Systematic Review

Abstract

Cloud computing has become a transformative paradigm, offering scalable resources, flexible storage, and on-demand services for modern applications. However, the rapid growth of cloud infrastructures has introduced challenges in joint resource allocation, efficient task scheduling, and data security. In large-scale environments, heterogeneous resources such as CPU, memory, bandwidth, and storage must be dynamically managed to ensure optimal performance while maintaining security. Traditional scheduling approaches often focus on single objectives like execution time or cost, overlooking complex resource interactions and security concerns. Recent advancements in deep learning–based optimization have enabled more intelligent resource management solutions. Techniques such as attention-based convolutional neural networks and pyramidal architectures can capture multi-scale features and improve adaptability. Hybrid models combining pyramidal convolution with split-attention mechanisms effectively learn hierarchical workload patterns, enabling accurate resource prediction, anomaly detection, and dynamic allocation. Additionally, the growing complexity of workloads from AI, IoT, and big data applications demands efficient strategies for load balancing, energy efficiency, latency reduction, and SLA compliance. Machine learning-driven scheduling frameworks enhance resource utilization and reduce operational costs, ultimately improving system performance and efficiency in cloud environments.