EnDASH-5G: Energy-Efficient Video Streaming over 5G
We have developed EnDASH-5G, a novel network-aware Adaptive Bitrate (ABR) streaming algorithm that optimizes UE energy efficiency for 5G mmWave networks. Our approach enhances video streaming performance while significantly reducing power consumption by synchronizing video data downloads with real-time network conditions.
Key Features
- Network-Aware Adaptation: EnDASH-5G leverages a transfer learning-based throughput prediction model to anticipate network variations.
- Energy-Efficient Bitrate Control: Uses Deep Reinforcement Learning (DRL) to dynamically adjust video bitrate and playback buffer length.
- Optimized for 5G mmWave: Evaluates real-time channel conditions to reduce RRC CONNECTED dwell time, lowering power consumption.
- Performance Gains:
- Achieves up to 30.5% energy savings compared to Pensieve ABR.
- Maintains near-optimal Quality of Experience (QoE).
- Validated via extensive ns-3 5G mmWave simulations.
Contributors
Abhijit Mondal
VDX TV
Argha Sen
IIT Kharagpur, India
Basabdatta Palit
NIT Rourkela, India
Sandip Chakraborty
IIT Kharagpur, India
Publications
- Basabdatta Palit, Argha Sen, Abhijit Mondal, Ayan Zunaid, Jay Jayatheerthan, Sandip Chakraborty: "Improving UE energy efficiency through network-aware video streaming over 5g", IEEE TNSM 2023
- Abhijit Mondal, Basabdatta Palit, Somesh Khandelia, Nibir Pal, Jay Jayatheerthan, Krishna Paul, Niloy Ganguly, and Sandip Chakraborty: "EnDASH-A Mobility Adapted Energy Efficient ABR Video Streaming for Cellular Networks", IFIP Networking 2020