Media on demand

In this paper we present a unified approach for delivering hypermedia/multimedia objects over broadband networks. The described service consists of several multimedia servers and a set of functions that intend to present to the end user interactive information in real-time. The hypermedia documents are structured using a hypermedia markup language that keeps information of the spatiotemporal relationships among document's media components. In order to deal with the variant network behavior, buffering manipulation mechanisms and grading of the transmitted media quality techniques are proposed to smooth presentation and synchronization anomalies. RÉSUMÉ: Dans ce document nous présentons une approche unifiée pour la fourniture d'objets hypermédias/multimédias sur des réseaux à larges bandes. Le service décrit est constitué de plusieurs serveurs multimédias ainsi que d'un ensemble de fonctions qui visent à présenter à l'utilisateur final une information interactive en temps réel. Les documents hypermédias sont structurés en utilisant un language "mark-up" qui garde l'information sur les liens spaciotemporels entre les différents composants des médias du document. Afin de faire façe à la réaction variable du réseau, des mécanismes de protection contre des erreurs de manipulation ainsi que des améliorations des qualités tecniques des médias transmis sont proposés. Ils visent à atténuer les anomalies de présentation et de synchronisation.

Distributed platforms for live and on-demand media streaming delivery such as content distribution networks and media ondemand systems, are being diffused mainly due to the widespread availability of IP-based, bandwidth-capable digital networks. Provision of multimedia group services is usually supported by transmitting media streams to subscribers organized in a multicast group. Although multicast streaming saves bandwidth and improves scalability, it is prone to be hacked. This paper proposes an efficient technique centered on the Blowfish symmetric encryption algorithm for securing media streams based on the Real-time Transport Protocol (RTP). The developed technique along with an ad-hoc key distribution mechanism is seamlessly embedded into our Java-based cooperative playback system-ViCRO C , which allows multicast transmission on-demand of archived multimedia sessions to a cooperative group of clients.

ith the proliferation of broadband access, the Internet faces the challenge of distributing rich media content to a worldwide audience of high-speed users. The challenges for streaming media today include the high data rates and significant bandwidth required for the uninterrupted delivery of high-quality music and video. Building a scalable and reliable system for on-demand and live streaming in this environment has proven to be very difficult. By definition, a media-on-demand system allows clients to access the media when they want and play it back without interruption after a given start-up latency. On our system, the outgoing bandwidth from the server is independent of the number of clients, and each client behaves independently of other clients. Furthermore, for each piece of media, the server minimizes the aggregate bandwidth requirement. Even though most codecs tolerate a certain amount of data loss, a high loss rate can significantly affect playback quality. Using an acknowledgment protocol like TCP to ensure reliability is extremely inefficient in one-to-many protocols like multicast. Requests for retransmission of lost packets can easily overwhelm a server, resulting in feedback implosion. Even with router assistance to aggregate acknowledgments, for a sufficiently heterogeneous network and a large enough audience, at least one client will lose almost every packet, requiring retransmission, which causes delays and results in massive reception of duplicate packets. An additional shortcoming with acknowledgment protocols is the delay associated with waiting for the lost packets to be retransmitted. Our MoD system solves these problems by using forward error correction codes to recover lost data. FEC uses the same redundant data to allow multiple clients to recover from different packet loss.