IA/0151 - QOE MANAGEMENT
Academic Year 2019/2020
Free text for the University
ALESSANDRO FLORIS (Tit.)
- Teaching style
- Lingua Insegnamento
|[70/91] INTERNET ENGINEERING||[91/00 - Ord. 2018] INGEGNERIA DELLE TECNOLOGIE PER INTERNET||3||30|
The QoE management laboratory has the objective to introduce the concepts of Quality of Experience and the most relevant QoE monitoring and management techniques for Internet of Things (IoT) systems and Smart City applications. Specifically, the topics are: utilization of sensing device and software applications to acquire information concerning the user and finalized to the estimation of the perceived QoE; statistical analysis for defining the most relevant QoE influence factor for the considered application; Machine Learning (ML) techniques for defining QoE prediction models; utilization of Apache Spark for data analysis and building of ML-based QoE prediction models.
Lexical: understanding and ability to use technical-scientific language
Informatics: ability to use / learn tools and software and basic Python programming skills.
Communications: knowing how to present concepts and information in oral, written and graphic form.
Organizational: know how to organize activities during the day and plan a medium-term work / study plan.
Knowledge: Quality of Service, Quality of Experience, Internet of Things, Machine Learning, Data analysis (short reviews will be made).
Skills: ability to define the link between phenomena, their properties and their abstract representation.
The laboratory includes the following activities:
1. Quality of Experience: definition of QoE; QoE evaluation methods; QoE models.
2. Internet of Things e Smart City: IoT definition and architecture; Smart City; Smart City applications and design techniques; QoE for IoT systems and Smart City applications.
3. QoE influence factors: definition and evaluation of QoE influence factors; environment sensing devices; software platform for acquiring QoE influence factors.
4. Data analysis: data analysis techniques; identification of the most relevant QoE influence factors.
5. Apache Spark: setup and basic functionalities; RDD creations and operation; interface with Hadoop and Cassandra; Machine Learning with Mllib.
6. QoE models: Machine Learning (ML) techniques; definition of predictive QoE models based on collected QoE influence factors; controls actions for QoE management based on predicted QoE.
The lab takes place during one semester; it is based on frontal lessons with presentations and group exercises. The laboratory has a total duration of 30 hours. During the lessons, participation is solicited through questions, requests for interpretation of the analytical results and reflections on the applicative aspects.
Verification of learning
Students will be asked to provide the results of a project assignment by the end of the semester. The project assignment will regard the study and design of a QoE management model for a specific application, including the needed experiments.
Lecture slides (EN)
Holden Karau, Andy Konwinski, Patrick Wendell & Matei Zaharia. Learning Spark. OReilly
Course slides in pdf format are provided.
The following open source tools are suggested: