Advanced search
Start date

Network measurement and modeling supporting resource allocation and user quality of experience

Grant number: 15/24434-5
Support type:Research Projects - Thematic Grants
Duration: October 01, 2017 - September 30, 2021
Field of knowledge:Physical Sciences and Mathematics - Computer Science
Principal Investigator:Edmundo Albuquerque de Souza e Silva
Grantee:Edmundo Albuquerque de Souza e Silva
Co-Principal Investigators:Daniel Sadoc Menasche ; DONALD FRED TOWSLEY ; Renata Cruz Teixeira ; Rosa Maria Meri Leão
Assoc. researchers: Carlos de Souza Lima Junior ; Celso Augusto Menezes de Andrade ; Diego Ximenes Mendes ; Fernando Carlos Azeredo Verissimo ; Gabriel Guimarães Braga dos Santos Mendonça ; Gaspare Giuliano Elias Bruno ; GUILHERME DA SILVA SENGES ; Gustavo Henrique Alves dos Santos


As society critically relies on the Internet, the network infrastructure and its impaction the Quality of Experience (QoE) perceived by users deserve more attention. Nowadays, the relationship between network characteristics and users QoE, although very fundamental, is still not well understood. The lack of measurements and of mathematical models to characterize users behavior are some among the numerous challenges faced by scientists and practitioners to better understandthe key factors that impact QoE. In this project, our goal is to model and analyze real data collected from Internet service providers (ISPs). The purpose of the proposed models is to issue online and offline recommendations about investments on network resources. To this aim, we will make use of a measurement environment created by TGR -- a startup which is our partner in this project -- in collaboration with NET. In addition, it is also our objective in this project to build a measurement testbed at COPPE (UFRJ). Our first envisioned contribution consists of the study and implementation of efficient policies to monitor the QoS and QoE, significantly scaling the monitoring that is already being conducted by TGR-COPPE-NET. As our second contribution, we intend to create probabilistic models from the collected data. Such models are helpful, for instance, to generate synthetic data. As our third contribution, we will build a framework that allows us to predict the QoE perceived by users, based on data collected in our measurements from the home users and from the access networks (corresponding to home users and ISP viewpoints, respectively). Finally, as our third contribution, we will parameterize analytical models and simulations of future internet environments, such as information centric networks, using collected data and the corresponding models. Our project embraces all the steps of the scientific method, from data collection up to model creation and validation. The results obtained through this project will be subject of transfer of technology to TGR and will be applicable to thousands of broadband users in Brazil through the TGR-COPPE-NET collaboration. (AU)