[Nfd-dev] Techniques for prioritizing audio in NdnCon

Burke, Jeff jburke at remap.UCLA.EDU
Sat Mar 28 12:01:56 PDT 2015



Hi folks,

We are trying to get NdnCon into use for practical conferencing soon.

It is clear that audio quality — in particular, minimizing dropouts — is critical.  Prioritizing audio over video performance when there are bandwidth / latency issues is something that can to some extent be done in the application.  But I am wondering if there are any tools that the architecture can provide as well.

In particular, can the notion of one-hop priority could be discussed further?

Originally, when we had discussed implementations of scalable video coding over NDN with Van, he had suggested that one-hop priority might be useful.  This is not end-to-end QoS but rather just a hint/request to the upstream hop on which interests the app would like fulfilled first.   It could be propagated or not.

In SVC streaming, such a capability would enable the consuming application to issue interests for the base layer with higher priority than those for the enhancement layer(s) without having to add complex machinery in the consumer to try to do this at the application level. (The pipelining/buffering machinery for low-latency communication is already somewhat tricky.)

Looking at ndnrtc now, this capability could actually be very useful for prioritizing interests audio over video as well.

What do people think?  Is this something that could be considered / experimented with?  Any other ideas?

Thanks,
Jeff

ps.  See my message to the ndn list – it would be great if NFD developers could try out NdnCon!




-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.lists.cs.ucla.edu/pipermail/nfd-dev/attachments/20150328/f89c0026/attachment.html>


More information about the Nfd-dev mailing list