| KEYWORD |
Optimal audio-mixing server placement for Networked Music Performance (NMP) frameworks
Parole chiave METAVERSE, MUSIC, OPTIMIZATION ALGORITHMS, STREAMING
Riferimenti CRISTINA EMMA MARGHERITA ROTTONDI
Gruppi di ricerca Telecommunication Networks Group
Descrizione Networked music performance (NMP) enables remote musicians to perform together by means of low-latency audio/video streaming over a telecommunication network. At present, no NMP application provides a suitable backend infrastructure that enables large scale deployments (e.g., orchestras and choirs) and is resilient to time-varying network traffic conditions. Available NMP applications either operate in peer-to-peer fashion or leverage single servers located at the premises of one of the performers or in institutions such as universities or music schools, without considering the geographical location of the users and with no flexibility on the placement of the server position.
The goal of this project is the design and implementation of optimization algorithms for the automatic placement of Virtual Machines (VMs) with running server instances to operate as backend infrastructure for the support of client/server-oriented audio/video streaming. The VMs will be possibly allocated in Edge Computing nodes located in the proximities of the participants in the NMP session. The optimization algorithms will permit to reduce the network delay component by exploiting traffic prediction methods based on Machine Learning to operate at various time scales, which will enable them to operate proactively and to dynamically adjust VM locations depending on the current and forecasted network congestion conditions. ML-based traffic prediction techniques will be also leveraged to proactively trigger adjustments in the audio/video streaming parameters with the aim of improving the Quality of Experience perceived by the users.
The thesis student will be co-tutored by prof. Luca Turchet (University of Trento).
Conoscenze richieste good programming skills, basic notions on operations research
Scadenza validita proposta 25/08/2024
PROPONI LA TUA CANDIDATURA