The Audio Cloud – Decreasing Cost Improving Reliability of Audio Links for Broadcast

摘要

Audio over IP is an accepted standard for audio delivery but there are further steps that can be taken to improve reliability and add greater redundancy. In this paper, I will set out four core techniques which, in various combinations, can create a type of audio ‘cloud’, a highly resilient system suitable for many different audio transport applications. The first technique is Redundant Streaming; the core on which any audio cloud is built. This is the ability to stream multiple copies of the same audio packet over divergent networks to protect against packet drop or loss of connections. Secondly, we explore Packet Forwarding or drop and insert on IP. This is the ability to make each decode site capable of supplying other decoders with audio packets. The third technique is Distributed Intelligence throughout the cloud; essentially the advanced control and monitoring functions needed to make the cloud function and self-heal with limited user intervention. The final component is that of relocating the Multicast or Multiple Unicast functionality away from the source encoder, reducing cost and improving reliability. The paper will examine both real world and theoretical scenarios based on the audio cloud components to produce highly resilient IP audio networks.