| Abstract |
Wireless access, use and traffic demand are on a fast rise. The number of mobile devices and their capabilities, accessing potentially multiple wireless network interfaces, also increase dramatically. Wireless networks experience “periods of severe impairment”, causing severe degradation to the performance of the service running on wireless devices and to the respective user experience. The crowd-sourcing and participatory sensing in large-scale systems provide the capability to monitor in real-time the quality of service (QoS).
Motivated by the need to enable users to select the appropriate provider/operator, especially when visiting a new environment, we developed the u-map system. A u-map client running on a smartphone uploads in regular basis information about the user profile (e.g., service requirements, device capabilities), QoE feedback about services (as indicated explicitly by the user), traffic demand statistics, network/spectrum conditions, and position of the device. This information is stored in a spatio-temporal geo-database, attached to a u-map server. A u-map client may query this database to obtain information about service providers and their customers’ satisfaction. In that way, a user can make an educated selection of the appropriate provider, when visiting a new area. The term “appropriate” denotes the provider that best matches the user’s criteria (e.g., it is the best in term of a certain QoS metric, or in QoE, in that area, or it supports specific features). Moreover, given that a smartphone may have access to several network interfaces simultaneously, a u-map client may select the appropriate wireless or cellular operator and network interface for offloading its traffic.
This thesis presents the u-map system and analyzes its performance, in terms of power consumption, responsiveness, and scalability. To understand how users assess such systems in general, and how they perceive the performance of the u-map and its services specifically, we also performed a 2-week field study with 21 real users. The outcomes of the performance analysis and the field study are encouraging: The power consumption and response delay of the u-map are relatively low. Most of the participants indicated that the u-map can improve their experience and helps in the discovery of networks that offer improved QoS. During the study, 30% of users discovered the best available network, improving substantially their experience. Users were also satisfied with the graphical user interface (GUI) of the u-map and other specific features, as well as the overall performance of the u-map service. Based on user feedback, we also modeled the user satisfaction about the u-map as a utility function. The model can predict the overall score of the user satisfaction about the u-map accurately. Finally, we applied the u-map in other domains, such as water distribution networks and medical applications, demonstrating its generality.
|
u-map: Empowering users and telecom providers by employing the crowd-sourcing paradigm
