Abstract

Cloud computing is an emerging paradigm for provisioning network, storage, and computing resources on demand using a pay-per-use model. Conferencing is the conversational exchange of media between several parties. Cloud-based conferencing services can provide benefits such as easy introduction of different types of conferences, resource usage efficiency and scalability.
A business model has been recently proposed in a position paper for cloud-based conferencing with the following roles: conference substrate provider, conference infrastructure provider, conference platform provider, conference service provider, and broker. Conference substrates are generally atomic and served as elementary building blocks (e.g. signaling, mixing) of conferencing applications. They can be virtualized and shared for resource efficiency purposes. Multiple conferencing substrates can be combined to build a conferencing service (e.g. a dial-out audio signaling conference service composed from dial-out signaling and audio mixer substrates).
The focus of this thesis is to design a semantic-oriented description framework for conferencing substrates and an architecture for their publication and discovery. The description framework is made up of a description language and a cloud-based conference ontology. The conference ontology is modeled on the basis of the interacting roles in the proposed cloud-based conferencing business model. The overall publication and discovery architecture for cloud-based conference substrates is made up of three brokers and the related publication and discovery interfaces. The publication and discovery interfaces are modelled using REpresentation State Transfer (REST) interfaces. A prototype is built to demonstrate the feasibility of this architecture. The effectiveness of the architecture is also proved using the performance measurements.