Optimized Resource Scheduling for Open Source Mobile Network Heterogeneous Computing Engines: A Comparative Study of DRF and H-DRF Algorithms

With the rapid development of the communications and computer industries, the demand for efficient data processing has surged. Single-processor architectures struggle to meet the high communication rate requirements of technologies such as V2X and AI, leading to a shift towards heterogeneous computing. Although heterogeneous computing improves data processing speed, it exacerbates the resource allocation problem. To address the shortcomings of the DRF scheduling method, this paper proposes an H-DRF scheduling approach based on DRF. Subsequently, it provides a conceptual framework for implementing H-DRF. To prevent issues such as project resource allocation failures and the allocation of critical resources, a weighting parameter is introduced to adjust the resource share ratios. An algorithmic approach is then outlined, and the H-DRF scheduling method is implemented in MATLAB. The performance of DRF and H-DRF is compared by allocating resources under identical system resources and project demands, leading to a conclusion. Finally, a comparative analysis of various parameters involved in the resource allocation process using DRF and H-DRF is conducted, explaining the superiority of H-DRF over DRF. Under the same conditions, H-DRF achieves higher resource utilization and better task throughput.