摘要：Third-generation (3G) cellular systems are often considered as a promising strategy for high-speed internet deployment in rural areas. However, bandwidth in 3G systems is a limited resource and can be quite different from the advertised bandwidth. In addition, in wireless applications such as voice over IP, video-conference, remote monitoring and telemetry (e.g., for healthcare applications) the Quality of Service (QoS) is critical. It is therefore important to understand how different factors affect the QoS requirements of real-time applications. Focusing on the reverse link (or uplink), we performed an assessment of the bandwidth and packet loss rate in a real network environment, using test data collected with two wireless data cards from different providers, in both rural and urban locations. The main contribution of this paper is the presentation of a formal statistical method – a Design of Experiments (DOE) analysis - that can be used to analyze the interactions between different variables, such as packet size, location, buffer size, and wireless provider. With a better understanding of the impact of such factors and their interactions, the end-user may be able to make the best selection of certain controllable parameters, in an effort to improve the QoS of the 3G connection. This is especially important for users located in rural areas, where cellular coverage is limited.
