SOSP '01 Proceedings of the eighteenth ACM symposium on Operating systems principles
Towards a Video QoE Definition in Converged Networks
ICDT '07 Proceedings of the Second International Conference on Digital Telecommunications
Resource Management Considerations in Collector Overlay Networks
ICNS '08 Proceedings of the Fourth International Conference on Networking and Services
Quality monitoring of video over a packet network
IEEE Transactions on Multimedia
Modeling packet-loss visibility in MPEG-2 video
IEEE Transactions on Multimedia
An Adaptive Media-Aware Retransmission Timeout Estimation Method for Low-Delay Packet Video
IEEE Transactions on Multimedia
| Hi-index | 0.00 |
To prevent subscriber churn, network service providers of VoD, SDV and IPTV have a pressing need to pro-actively detect, isolate and fix outages within an access network. Network induced degradations prove to be detrimental for streaming applications. This typically leads to a poor quality of experience (QoE) for subscribers. By monitoring key functional points of the access network for traces of degradation, service providers can devise mechanisms to mitigate the problem. In this work we propose a hierarchy of exporters, collectors and ANCON (ANalysis and CONtrol) nodes that can semiautonomously monitor, detect and isolate impairments within an access network. Exporters on the data plane gather and disseminate statistics for individual subnets, which are streamed onto "collector" nodes on an orthogonal plane. Collector nodes aggregate traffic from various exporters, and stream them onto the root of the tree (ANCON). With an even placement of exporters, root cause analysis can now take the granularity of loss rates or delay rates in individual segments or subnets of an access network. As an extension to our architecture, we show that the overlay can support instrumentations of quality evaluation for streaming video. As an example, we use a simple MOS model that is in part an extension of the ITU-T Erlang model to predict the quality of a video stream much before it reaches the end user. Extensive simulations are presented to justify the design choices made in the process.