LSTM-Based DWBA Prediction for Tactile Applications in Optical Access Network

Historically, the optical access network (OAN) plays a crucial role of supporting emerging new services such as 4 k, 8 k multimedia streaming, telesurgery, augmented reality (AR), and virtual (VR) applications in context Tactile Internet (TI). In order to prevent losing connectivity current mobile Internet, OAN must expand capacity improve quality Services (QoS) mainly for low latency 1 ms. The has adopted artificial intelligence (AI) technology, deep learning (DL), classify predict complex data. This trend focuses on bandwidth prediction. software-defined (SDN) cloud technologies provide all essential capabilities deploying enhance performance next-generation ethernet passive networks (NG-EPONs). Therefore, this paper, we propose long-short-term-memory model-based predictive dynamic wavelength allocation (DWBA) mechanism, termed LSTM-DWBA NG-EPON. Future end-user is predicted based NG-EPON MPCP control messages exchanged between OLT ONUs cycle times. proposed addresses uplink message overhead QoS bottleneck networks. Finally, extensive simulation results show packet delay, jitter, drop, utilization.