الملخص الإنجليزي
Abstract
Recent breakthroughs in cellular network technologies lead to a rapid increase in the multimedia usage as different types of services are supported. Monitoring the cellular networks is a necessity in order to optimize network operations to provide high quality services to subscribers. Traditionally, Quality of Service (QoS) has been used to assess network performance from technical point of view, where network metrics are gathered and compared with a threshold to guarantee that the network is operating as expected. However, it is a challenge for the service providers to guarantee users. satisfaction with the services they use. Moreover, QoS performance metrics are reported as absolute values, where these metrics are a representation of the network performance rather than a specific service. As more subscribers join the network, it is essential for service providers to measure acceptability of a service from the users' perspective to ensure high quality experience. Hence, Quality of Experience (COE) is proposed as an alternative. In this thesis, QoE of data services are estimated using a Deterministic QoE model (DQX), which represents the QoE as Mean Opinion Score (MOS). The model evaluates MOS for different services individually, using network parameters that can be measured from any handset. In addition, the thesis rely on crowdsourcing to measure the network metrics. This is opposed to what is currently followed by operators, where QoS measurements are conducted through field tests that are time consuming and less representative of the user's experience. As such, a QoS monitoring
App has been further customized to measure the QoS parameters for selected services (i.e. web browsing, social networking, and video streaming). These measurements were conducted in Muscat using cellular network of two operators. Furthermore, the DQX model was simulated using MATLAB model that was developed in this thesis. The MATLAB model uses the measured metrics as input. The results obtained in this thesis generate the MOS of each service, giving the operator and regulator an insight of how users perceive the network performance.
