INDIGO: A Generalized Model and Framework for Performance Prediction of Data Dissemination
Staff - Faculty of Informatics
You are cordially invited to attend the PhD Dissertation Defense of Kamini GARG on Monday, February 27th 2017 at 14h30 in room SI-006 (Informatics building)
According to recent studies, an enormous rise in location-based mobile services is expected in future. People are interested in getting and acting on the localized information retrieved from their vicinity like local events, shopping offers, local food, etc. These studies also suggested that local businesses intend to maximize the reach of their localized offers/advertisements by pushing them to the maximum number of interested people. The scope of such localized services can be augmented by leveraging the capabilities of smartphones through the dissemination of such information to other interested people.
To enable local businesses (or publishers) of localized services to take informed decision and assess the performance of their dissemination-based localized services in advance, we need to predict the performance of data dissemination in complex real-world scenarios. Some of the questions relevant to publishers could be the maximum time required to disseminate information, best relays to maximize information dissemination etc. This thesis addresses these questions and provides a unified solution called INDIGO that enables the prediction of data dissemination performance by collectively considering the real-world aspects of data dissemination process under different variations of physical and social proximity among people.
INDIGO empowers publishers to assess the performance of their localized dissemination based services in advance both in physical as well as the online social world. It provides a solution called INDIGO–Physical for the cases where physical proximity plays the fundamental role and enables the tighter prediction of data dissemination time and prediction of best relays under real-world mobility, communication and data dissemination strategy aspects. Further, this thesis also contributes in providing the performance prediction of data dissemination in large-scale online social networks where the social proximity is prominent using INDIGO–OSN part of the INDIGO framework.
INDIGO is the first work that provides a unified solution and enables publishers to predict the performance of their localized dissemination based services under different variations of physical and social proximity among people and different real-world aspects of data dissemination process. INDIGO outperforms the existing works for physical proximity by providing 5 times tighter upper bound of data dissemination time under real-world data dissemination aspects. Further, for social proximity, INDIGO is able to predict the data dissemination with 90% accuracy and differently from other works, it also provides the trade-off between high prediction accuracy and privacy by introducing the feature planes from online social networks.
- Prof. Mehdi Jazayeri, Università della Svizzera italiana, Switzerland (Research Advisor)
- Prof. Silvia Giordano, University of Applied Sciences and Arts of Southern Switzerland, Switzerland (Research co-Advisor)
- Prof. Fabio Crestani, Università della Svizzera italiana, Switzerland (Internal Member)
- Prof. Cesare Pautasso, Università della Svizzera italiana, Switzerland (Internal Member)
- Prof. Sajal Das, Missouri University of Science and Technology, USA (External Member)
- Prof. Oscar Mayora, CREATE-NET Trento, Italy (External Member)