HARDWARE–SOFTWARE PLATFORM ARCHITECTURE FOR CLOUD COMPUTING ACCELERATION USING FPGA

Abstract

The increasing computational load in cloud infrastructures driven by large-scale data analytics and machine-learning tasks requires efficient hardware acceleration solutions. This paper presents a hybrid hardware–software architecture integrating FPGA-based accelerators into a cloud platform for data-intensive and inference-driven applications. The proposed approach combines reconfigurable hardware logic with containerized software environments, achieving 3–5× improvement in performance and up to a 40% reduction in energy consumption compared to CPU-based systems. Experimental evaluation shows that the FPGA-enabled architecture provides scalable, low-latency execution suitable for high-throughput workloads in modern data centers.