Traffic estimation based on long short-term memory neural network for mobile front-haul with XG-PON Download: 771次
1 Key Laboratory of Optical Fiber Sensing and Communications, Ministry of Education, University of Electronic Science and Technology of China, Chengdu 611731, China
2 Business School, University of International Business and Economics, Beijing 100029, China
Figures & Tables
Fig. 1. MFH architecture based on the XG-PON system.
下载图片 查看原文
Fig. 2. LSTM neural network architecture used in the proposed method.
下载图片 查看原文
Fig. 3. LSTM cell and its unfolding in time.
下载图片 查看原文
Fig. 4. Detailed structure of the LSTM cell memory block. The sigmoid function is denoted as and the pointwise multiplication is denoted as in the figure.
下载图片 查看原文
Fig. 5. MSE in the training process.
下载图片 查看原文
Fig. 6. Upstream delay performance comparison of RR-DBA, FNN-DBA, and LSTM-DBA.
下载图片 查看原文
Fig. 7. Upstream jitter performance comparison of RR-DBA, FNN-DBA, and LSTM-DBA.
下载图片 查看原文
Fig. 8. Upstream packet loss ratio performance comparison of RR-DBA, FNN-DBA, and LSTM-DBA.
下载图片 查看原文
Fig. 9. Upstream delay performance comparison of RR-DBA, FNN-DBA, LSTM-DBA, and FBA for one active ONU case.
下载图片 查看原文
Table1. Traffic Simulation Parameters
Parameters | Values |
---|
Number of bursts | 5000 | Mean burst time length | 2 ms | Poisson arrival rate | [95, 110, 125–200 Mbps] | Hurst parameter | 0.8 | Pareto shape parameter | 1.4 | Packet size | 1470 bytes |
|
查看原文
Table2. DBA Simulation Parameters
Parameters | Values |
---|
Application traffic model | PPBP | Simulation time | 10 s | Max polling interval (all DBAs) | 125 μs | Number of RRHs (ONUs) | 10 | T-CONT per ONU | 1 | Roundtrip propagation delay | 100 μs | ONU queue size (T-CONT buffer) | 1 Mbytes |
|
查看原文
Min Zhang, Bo Xu, Xiaoyun Li, Yi Cai, Baojian Wu, Kun Qiu. Traffic estimation based on long short-term memory neural network for mobile front-haul with XG-PON[J]. Chinese Optics Letters, 2019, 17(7): 070603.