RT info:eu-repo/semantics/article
T1 Green PON Deployment: Enhancing Performance and Sustainability Using SLA-Aware DBWA Algorithms for 25G/50G-EPON
A1 Garg, Sukriti
A1 Merayo Álvarez, Noemí
K1 50G passive optical networks
K1 dynamic bandwidth and wavelength allocation
K1 emerging services
K1 energy savings
K1 service level agreements
AB Emerging applications such as extreme virtual reality, enterprise access, 5G front-haul / back-haul, and future 6G immersive services demand exceptional bandwidth and low latency. Next-generation passive optical networks (PONs), including 25G/50G-EPON (Ethernet PON) and 50G-PON, address these requirements. However, the telecoms sector faces sustainability challenges, with networks consuming over 75% of the sector’s energy and contributing more than 25% of greenhouse gas emissions, which is increasing with bandwidth demand. Fiber-based PON networks offer high-speed, low-latency, and energy-efficient solutions. To further enhance energy efficiency, we developed a dynamic bandwidth wavelength allocation (DBWA) algorithm for 50G-EPON systems, adapted for single- and dual-wavelength configurations. For single-wavelength ( 1×50 G) configurations, the sleep mode aware algorithm (SMA) activates the optical network unit (ONU) sleep mode when no data are transmitted, reducing energy consumption and ensuring minimal delay due to the absence of wavelength switching. For dual-wavelength ( 2×25 G) systems, the online restricted wavelength switching (RWS) algorithm dynamically allocates wavelengths upon receiving user reports, regulating the switching to minimize delays while achieving energy savings at the ONU. Finally, leveraging a predictive approach, the offline restricted wavelength usage and switching (RWUS) algorithm evaluates bandwidth demands over 0.1-second intervals, enabling energy-efficient wavelength switching and optimizing both optical line terminal (OLT) and ONU energy usage. The simulation results highlight the superior energy efficiency of the offline RWUS algorithm and the extended ONU sleep times, especially under low load conditions, while maintaining the lowest delay among multi-wavelength algorithms. Conversely, despite being less energy-efficient, the SMA algorithm scores slightly better in delay for single-wavelength systems. All algorithms ensure compliance with service level agreements (SLAs), providing reliable bandwidth allocation for various profiles. These results support green PON deployments and recommend the RWUS algorithm for two-wavelength 50G-EPON systems ( 2×25 G) and SMA for single-wavelength 50G-EPON ( 1×50 G) to optimize energy efficiency, latency, and quality of service.
PB IEEE
SN 2169-3536
YR 2025
FD 2025
LK https://uvadoc.uva.es/handle/10324/80129
UL https://uvadoc.uva.es/handle/10324/80129
LA eng
NO IEEE Access, 2025, vol. 13, pp. 158028-158043
NO Producción Científica
DS UVaDOC
RD 28-nov-2025