THESIS
2000
90 leaves : ill. ; 30 cm
Abstract
In this thesis we develop the architectural designs of our Intelligent streaming Gateway (ISG) as well as a distributed network for large-scale live streaming broadcast over the Internet. Our contributions solve the problems in network and system scalability by our philosophy of "pushing the content distribution and duplication at the edge of network". In particular, we concentrate on showing how the novelty of ISG can successfully achieve this philosophy, and improve the overall broadcasting quality and performance. We have four niche mechanisms developed as the architectures for a live streaming system and content delivery network - smart streaming consolidation, high performance streams duplication, distributed dynamic load-balancing tree protocol and adaptive inter-domain streaming...[
Read more ]
In this thesis we develop the architectural designs of our Intelligent streaming Gateway (ISG) as well as a distributed network for large-scale live streaming broadcast over the Internet. Our contributions solve the problems in network and system scalability by our philosophy of "pushing the content distribution and duplication at the edge of network". In particular, we concentrate on showing how the novelty of ISG can successfully achieve this philosophy, and improve the overall broadcasting quality and performance. We have four niche mechanisms developed as the architectures for a live streaming system and content delivery network - smart streaming consolidation, high performance streams duplication, distributed dynamic load-balancing tree protocol and adaptive inter-domain streaming distribution. All of them can be practically implemented as a real system, which is robust under rapidly changing environments such as the Internet. An innovative and critical object-oriented application programming interface (API) - Pipe object, is designed to make these four niches possible. We focus on using rigorous time complexity analysis, case analysis and experimental tests to justify the proposed architectures so that they may be applied in real systems with greater confidence.
Post a Comment