Efficiency, quality, and robustness of multimedia delivery over best-effort packet networks can be improved by advanced streaming protocols. This comes, however, at the cost of increased application complexity because existing solutions require close cooperation between transport layers and applications to perform effective adaptations. Media encoders have to become network-aware while transport layers need to consider certain properties of the data they deliver, such as the unequal importance of data units and their timing constraints. Today's adaptive streaming systems are developed jointly, customising encoders, protocols and surrounding code for a specific application. This heavily restricts the reuse of protocol implementations for other encoding formats, different application requirements and network environments.
In order to enable reuse and simplify the integration of advanced streaming protocols into applications, we propose the Noja programming model for multimedia middleware layers. Noja's abstractions leverage protocol coordination and provide selective transparency features to effectively hide complexity, but expose necessary control options to application developers. Based on examples we show that Noja is applicable to a broad range of multimedia streaming applications and beyond that to data streaming applications in general.
可以通过先进的流协议来提高通过尽力而为的分组网络进行多媒体交付的效率,质量和健壮性。然而,这是以增加应用程序复杂性为代价的,因为现有解决方案需要传输层和应用程序之间紧密合作以执行有效的适配。媒体编码器必须具有网络感知能力,而传输层则需要考虑它们传递的数据的某些属性,例如数据单元的重要性和时序限制等。当今的自适应流媒体系统是联合开发的,可针对特定应用定制编码器,协议和周围的代码。这严重限制了协议实现对其他编码格式,不同的应用程序要求和网络环境的重用。 P>
为了实现重用并简化高级流协议到应用程序的集成,我们提出了用于多媒体中间件层的Noja编程模型。 Noja的抽象利用协议协调并提供选择性的透明功能来有效隐藏复杂性,但向应用程序开发人员提供必要的控制选项。从实例中可以看出,Noja不仅适用于广泛的多媒体流应用程序,而且还适用于一般的数据流应用程序。 P>
机译:时钟层次结构:用于分组和控制媒体流的抽象
机译:使用跨层优化流控制传输协议的媒体独立切换传输
机译:无线/有线分组网络中多媒体流的跨层错误控制
机译:跨层控制媒体流的中间件抽象
机译:通过3G无线网络进行视频流的跨层拥塞控制
机译:大数据平台上异构数据实时分析的分布式流处理中间件框架:环境监测案例
机译:时钟层次结构:用于分组和控制媒体流的抽象