Innovative Approaches for Network Analysis and Optimization: Leveraging Deep Learning and Programmable Hardware

Autor(es)

de Castro, Ariel Goes ; Rothenberg, Christian Esteve

• Tipo de documento: Artigo Científico
• Fonte: 2024 IEEE 10TH INTERNATIONAL CONFERENCE ON NETWORK SOFTWARIZATION, NETSOFT 2024; v. N/A, p. 4-pg., 2024-01-01.

Resumo

Network demand for real-time applications like self-driving cars and cloud gaming strains existing networks. Latency and congestion hurt user experience. Realistic testing is vital to improving networks, but real-world data is scarce. In this context, we propose to analyze existing network data and identify traffic patterns and anomalies. We believe this knowledge can be used to feed generative adversarial network (GAN) models, which can create realistic synthetic data, supplementing existing real traces while protecting end-user privacy. This augmented data can then be used, for instance, to empower improved routing algorithms designed to benefit from programmable hardware (e.g., SmartNICs) and collected data plane metrics, paving the way for improved network performance and enhanced user experience with more autonomous decisions. This paper presents our initial analysis of synthetic network data generation technologies and summarizes the main ideas guiding my Ph. D. research. (AU)

• Processo FAPESP: 21/00199-8 - Redes e serviços inteligentes rumo 2030 (SMARTNESS)
• Beneficiário: Christian Rodolfo Esteve Rothenberg
• Modalidade de apoio: Auxílio à Pesquisa - Programa Centros de Pesquisa em Engenharia