Unveiling Performance Evaluation Unmasking NFV Infrastructure Against Traditional Hardware Networks Using TOPSIS Method
REST Journal on Data Analytics and Artificial Intelligence · 2025
The swift advancement of networking technologies has given rise to Network Function Virtualization (NFV) as a promising avenue for enriching network adaptability, scalability, and cost efficiency. This article delves deeply into the essential endeavour of appraising the effectiveness of NFV infrastructure and drawing a parallel with conventional hardware-based networks. Through the utilization of the Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) method, the primary objective of this study is to identify possible bottlenecks and prospects for enhancement within both approaches. In recent times, NFV has garnered substantial attention owing to its capability to uncouple network functions from dedicated hardware, allowing their operation within virtualized environments. While this shift offers agility and economic advantages, it simultaneously introduces intricacies necessitating meticulous performance evaluation. On the other hand, traditional hardware-based networks, while tried and assessed, may encounter constraints in promptly acclimating to dynamic requisites. As a result, an all-encompassing assessment framework becomes indispensable to accurately assess the genuine potential of NFV infrastructure. A robust research approach is utilized in this study, characterized by a quantitative analysis framework. Key performance indicators such as latency, throughput, resource utilization, and scalability are assessed and contrasted between Network Function Virtualization (NFV) infrastructure and conventional hardware networks. The Technique for Order of Preference by Similarity to Ideal Solution (TOPSIS) methodology is implemented to pinpoint performance limitations and potential enhancement zones within both paradigms. Evaluation and Alternate parameters taken as in this study, Scalability, Resource Utilization, Latency, and Energy Consumption serve as assessment parameters, while contrasting Traditional Hardware, NFV with Software Acceleration, NFV with GPU Acceleration, NFV with FPGA Acceleration, and Cloud-based NFV configurations. The results presented in the evaluation demonstrate the effectiveness of different Network Function Virtualization (NFV) alternatives using the TOPSIS method. NFV with GPU Acceleration emerges as the top-performing choice, closely followed by NFV with FPGA Acceleration. These alternatives exhibit superior proximity to the ideal solution, as indicated by their high Closeness Coefficient (Ci) values and top ranks. NFV with Software Acceleration ranks third, showcasing favorable performance. Cloud-based NFV and Traditional Hardware secure lower ranks, implying comparatively less optimal performance. The Closeness Coefficient values and ranks provide a clear and concise basis for decision-making, guiding the selection of NFV alternatives with greater alignment to desired outcomes.