A Novel Adaptive QoS Guarantee Framework for SMIL Multimedia Presentations with ABR Service

• Main Authors: Sheng-Wen HSIEH, 謝盛文
• Other Authors: Ing-Chau Chang
• Format: Others
• Language: en_US
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/11627983528683456230

碩士 === 朝陽科技大學 === 資訊管理系碩士班 === 89 === With technical advances of networks and computers in recent years, the multimedia services are more and more popular. Digital cameras and video recorders let us record what we see and hear easily in our daily life. The Synchronized Multimedia Integration Language (SMIL) and Asynchronous Transfer Mode (ATM) will make the transmission and playback of synchronous multimedia presentation more conveniently and easily. However, there are issues need to be overcome. For example, network transmission supports will affect the playback quality, which is the key factor for successful multimedia presentations. The playback quality is usually called as Quality of Service (QoS) for multimedia presentation. Different users may need different multimedia services and different multimedia services may require different QoS supports. In this thesis, we will analyze the SMIL document and dynamically adjust network bandwidth allocations among current media objects in the SMIL document to limit the transmission delay. In this way, playbacks of the SMIL multimedia objects in the SMIL presentation are more smoothly, which in turn guarantees the presentation QoS of the SMIL documents. We will propose a new SMIL QoS guarantee framework architecture in this thesis. This architecture considers requirements of the SMIL application and control mechanisms of the ATM network at the same time. This new architecture monitors the supported QoS (bandwidth, jitter, fairness) of the ATM network and the QoS of the SMIL presentation, which is called as Application Quality of Service (APQoS) in this thesis. If the network cannot satisfy the APQoS for the SMIL presentation, it will adjust the flow control of the ATM ABR service. This architecture includes the Multimedia Control Layer (MCL), the QoS Guarantee Layer (QGL) and the Flow Control Layer (FCL). The MCL will parse the SMIL file to retrieve useful information, such as the APQoS and Media Importance (MI). The QCL will adjust bandwidth allocations among media objects simultaneously according to the information from the MCL and the FCL. The FCL will use the technique of the neural network to let the ATM ABR service archive the high utilization, low cell loss rate and the bandwidth fairness. With this architecture, we can the guarantee QoS of the SMIL presentation with the ATM ABR service.