Architectures for Video-on-Demand content distribution in managed networks

摘要: La demanda de contenidos video ha aumentado rapidamente en los ultimos anos como resultado del gran despliegue la TV sobre IP (IPTV) y variedad servicios ofrecidos por operadores red. Uno que se vuelto especialmente atractivo para clientes es el bajo (VoD) tiempo real, ya ofrece una transmision (streaming) inmediata video. El precio tienen pagar este servicio aumento trafico las redes, estan cada vez mas congestionadas debido a mayor VoD al calidad propios Asi, uno principales objetivos esta tesis encontrar soluciones reduzcan nucleo red, manteniendo nivel adecuado reduciendo coste trafico. propone un sistema jerarquico servidores streaming ejecuta algoritmo ubicacion optima acuerdo con comportamiento usuarios estado Debido cualquier optimo distribucion alcanza limite no puede llegar nuevas mejoras, inclusion (los peers) proceso reducir aun Este logra aprovechando control operador tiene redes gestion privada equipos receptores (Set-Top Box) ubicados instalaciones clientes. reserva cierta capacidad almacenamiento peers almacenar transmitirlos otros fin aliviar streaming. incapacidad sustituir completamente streaming, asistido peers. Algunas cuestiones importantes abordan son saber parametros distintas distribuciones afectan rendimiento general sistema. Para dar respuesta estas preguntas, modelo estocastico preciso flexible cuenta capacidades enlace subida peers, numero tamano biblioteca video, esquema estimar beneficios trabajo tambien version extendida matematico mediante probabilidad fallo su recuperacion conjunto modelo. Estos modelos utilizan herramienta realizacion exhaustivos analisis amplia gama definidos modelos. Abstract The demand of contents has rapidly increased in the past years as result wide deployment IPTV and variety services offered by network operators. One that especially become attractive to customers is real-time Video on Demand because it offers an immediate large contents. price operators have pay for this convenience traffic networks, which are becoming more congested due higher quality videos. Therefore, one main objectives thesis finding solutions would reduce core network, keeping service satisfactory level reducing cost. proposes system hierarchical structure servers runs algorithm optimal placement according users’ behavior state network. Since any content distribution reaches limit upon further improvements can be made, including themselves (the process traffic. This achieved taking advantage operator privately managed networks over Set-Top Boxes placed at clients’ premises. reserves certain storage capacity store stream them other clients order alleviate servers. Because inability completely substitute servers, peer-assisted Some important questions addressed how parameters various distributions impact overall performance. In give answers these questions, precise stochastic model takes into consideration like uplink number size library, scheme estimate benefits work also extended mathematical failure probability their recovery time set parameters. These models used tools conducting thorough analyses range defined