User-centric video streaming quality of service and experience evaluation over next generation mobile networks

Abstract

Quality of Experience (QoE) is a prototype of human feelings, wishes, impressions and motives. Quality of Service (QoS) on the other hand is a technology-based metric that is employed to gauge how a network performs at any given time. Latterly, QoE in Next Generation Mobile Networks ( GMNs) - mostly referred to as 4G Long Term Evolution (LTE) networks have continued to attract interest from network researchers, developers and designers. This is mainly because QoE on its own comes with a degree of relationship to QoS or network conditions, yet to end users QoE matters much more than QoS. QoS is basically about network metrics such as jitter, pack delay and loss. Much as these metrics are critical too, they are only a small part of satisfying users, this is because QoE is more related to the perception of a service by a mobile user. The main challenge that remains is; mobile users always expect their mobile devices to have full connectivity and high network performance, if any of this is absent, providers of this service do not have sufficient time to wait for these quality complaints from the users to come through; instead such users will simply change service providers in the hope of finding a better service on another network. This confirms that excellent network metrics can easily mean nothing especially if users decide to find other service providers. This occurrence pauses two challenges; first, mobile users will continue to look for better quality video content streaming without breakouts elsewhere, secondly the networks they may keep moving to are bound to experience a knock-on effect as far as network resource utilization is concerned, this is mainly because of a sudden increase in the number of users accessing the same service at the time. In this same period, video streaming also continues to have a significant influence on network traffic especially in NGMNs. The major force behind this influence has been satisfaction of user experience as far as service delivery is concerned. Therefore, it is important that service providers hold a means of continuity in the measurement of the QoE since QoE may not only be judged on network speed improvement such as deploying advanced technologies. The notion of bridging this gap between QoS and QoE in NGMNs has indeed led to significant and successful works being published by network researchers. However, much of these systems and mechanisms suggested in these works mainly focus on presenting specific technologies. In this thesis, we introduce an independent mechanism that is compatible enough with other technologies. The mechanism has the ability

to monitor network QoS and at the same time evaluate trends in QoE as far as user perception of the service is concerned. We used both subjective lab environment simulations and a simulated environment using the NETEM simulator. To generate the video streams we used the VLC media player, a router and a switch hosted the Network Address Translator (NAT) on the sub-net and a tool known as Distributed Internet Traffic Generator (DITG) was used to emulate TCP bulk traffic. To emulate packet loss, Netem was our tool of choice, which was integrated into the Linux iproute2 tool set that we also used forthe queuing disciplines required to implement and realize the required behaviour. After this setup, we then ran several simulations in order to study the performance of our mechanism. According to the results, the mechanism performed as expected and each of the simulations resulted in improved quality for the relevant users. It should be noted that, customer differentiation presented the most expected results, which was an important aspect to Internet Service Providers(ISPs). The simulations also showed that the perceived quality of the users can be improved without sacrificing the quality of streams belonging to ordinary users. The general conclusion from the implementation of these experiments indicated that its indeed possible to monitor QoE of the users so as to retain them, this was entirely made possible because the mechanism that was implemented can equally keep track and evaluate network QoS in terms of video quality performance, resource utilization and optimization. Thus addressing the major challenges that QoS and QoE pause to NGMNs as the number of mobile video streaming users increase for different reasons.