dc.description.abstract
The audio-visual media content production industry (e.g.: broadcasters, small production companies) has been, and already is, employing rigid and difficult to scale technologies to transport and manage their streams through their processing chain. Although since early 2000s a gradually adoption of IP technologies has been happening, the process is still slow. Furthermore, most of the existing technologies involve large deployment and maintenance costs (e.g.: specific hardware, specific and costly wiring). For this reason, the study of IP technology is proposed, specifically technology related to the distributed cloud computing concept, in order to propose solutions to reduce costs and increase the audiovisual content production's possibilities. Particularly, this thesis focuses on analysing, proposing, developing and demonstrating specific virtualization, monitoring and application technologies in order to provide solutions to these mentioned issues. Regarding virtualization, technologies based on Linux Containers are used, specifically Docker containers. Thanks to the managing layer offered by Docker containers the shipment, distribution and execution of applications over the network is achieved. Moreover, platform scalability is assured because the maintenance, management and replication of applications containerized within this technology are fast and reliable. The use of monitoring tools is a key point to offer application status management to applications and to the platform itself and to allow the application of actuation policies in real-time in an efficient manner. Specifically, Collectd and Graphite are the selected tools. Moreover, these tools are able to be managed inside containers in order to be simultaneously deployed over the network together with the applications' platform. Finally, as it is demonstrated, the core of the platform, the LiveMediaStreamer framework, achieves the requirements in order to be used as a real-time cloud service for audiovisual media content production. This is thanks to the technologies above-mentioned, the statistics layer implemented for monitoring (network and performance) and the development of a REST API middleware.
