Workload-Aware Resource Profiling and Dynamic Energy Characterization in Multi-Tier Cloud Architectures

Cloud data centers increasingly host diverse applications with varying resource demands, making workload behavior a critical factor in achieving energy-efficient operation. Modern cloud infrastructures are typically multi-tier and heterogeneous, consisting of servers with different performance capabilities and power characteristics. While prior research has addressed energy-aware resource management, limited attention has been given to understanding how distinct workload patterns dynamically influence energy consumption across multiple architectural tiers. This research addresses the problem of inefficient energy utilization caused by static or workload-agnostic resource provisioning in cloud environments. Such approaches fail to capture the temporal and behavioral variations of CPU-, memory-, and I/O-intensive workloads, leading to suboptimal energy performance and resource underutilization. To overcome this limitation, the study proposes a workload-aware resource profiling framework that dynamically characterizes workload behavior and maps it to energy consumption patterns in multi-tier cloud architectures. The framework classifies workloads based on utilization signatures and correlates them with tier-specific energy profiles, enabling adaptive resource-aware decision-making. Dynamic energy characterization models are developed to reflect real-time workload transitions and infrastructure heterogeneity. The expected outcomes of this research include improved accuracy in energy estimation, enhanced workload-to-resource matching, and reduced overall power consumption without compromising system performance. The proposed framework provides a foundational intelligence layer that supports adaptive scheduling, energy-aware optimization, and sustainable cloud infrastructure design.