دانلود رایگان مقاله تخصیص منابع هوشمند برای بهبود QoE در زیرسیستم های چند رسانه ای IP

ABSTRACT

IP Multimedia Subsystem (IMS) is a robust multimedia service. IMS becomes more important when delivering multimedia services. Multimedia service providers can benefit from IMS to ensure a good QoE (Quality of Experience) to their customers with minimal resources usage. In this paper, we propose an intelligent media distribution IMS system architecture for delivering video streaming. The system is based primarily on uploading a multimedia file to a server in the IMS. Later, other users can download the uploaded multimedia file from the IMS. In the system, we also provide the design of the heuristic decision methods and models based on probability distributions. Thus, our system takes into account the network parameters such as bandwidth, jitter, delay and packet loss that influence the QoE of the end-users. Moreover, we have considered the other parameters of the energy consumption such as CPU, RAM, temperature and number connected users that impact the result of the QoE. All these parameters are considered as input to our proposal management system. The measurements taken from the real test bench show the real performance and demonstrate the success of the system about ensuring the upload speed of the multimedia file, guaranteeing the QoE of end users and improving the energy efficiency of the IMS.

7. Conclusion and future work

In this paper, we presented an architecture for the IMS system in order to deliver multimedia streaming. The system covered three aspects, the multimedia upload clients, the system manager, and the download clients. The proposed algorithm in the system manager is based on the intelligent media distribution, which decided how the upload user can select the optimal server to upload the media file and then how the group of users can receive the message regarding the updated contents in the IMS system. Accordingly, by taking effective of QoS parameters and the characteristics of the video on the system performance, we described some resource parameters to evaluate the performance regarding the energy consumption and the QoE. The evaluation results depicted that, the increase of the number of connected users had influenced on the performance of the resources such as rate of the CPU usage and the system temperature. In the experiments, we observed that the impact of transcoding is higher especially when the dynamic video was transcoded. Further, we evaluated the QoE in terms of relevant measures such as frame loss, subjective, and objective perceived video quality. The results concluded that MP4 used higher bitrate than XVID and rate of the dropped frames of MPEG4 was 65%, which was higher than the rate of H264. The rate of the i-frame of H264 was higher than MPEG4 and XVID for the high bitrate. MPEG4 presented higher PSNR than XVID, and XVID presented better MOS than other codecs.

In future work, we will use a sophisticated approach based on deep learning in order to increase the number of parameters which decide on the performance of the system and QoE. Therefore, we will work on reducing the effects of competing among clients, which leads to improvement of the perceived video quality.