On 2 January 2013 13:27, Dan Reardon <[email protected]> wrote: > (a) Has anyone used PGM over WIFI to devices yet? Anyone tried it with > hundreds or thousands of phones? (b) How well does it (or should it) work? >
WiFi always has scalability issues, some intelligence is probably requiring deploying to large scale across many access points. > (c) If the wifi / electrical environment is dirty, will zeroMQ's queuing > mechanisms get in my way? > Yes. ZeroMQ is asynchronous, it's about dumping data as fast as possible to the destination, for audio you want low-latency unreliable messaging requiring no round trips. > "guaranteed delivery", as this is a real time broadcast, missing audio > segments should be substituted with silence when the cannot > Spotify handles this a little better with white noise. There is an option of FEC to improve quality in lossy environments. > Is there any chance that anyone would want to implement Uflood, Ucast or > any other protocol specifically designed for wifi? > WiFi generally means completely random data-loss due to radio interference, wired paths suffer more predictable data-loss and typically burst outages. > We may be in a position to provide a testbed environment for large-scale > wifi testing if someone else wants to develop the code! Who should we be > talking to? Anyone interested in being the lead archtect or developer if we > can support testing? > I actually have some hardware that may be suitable for such testing, 29 small devices with limited CPU and memory that can be networked via radio or cable. -- Steve-o
_______________________________________________ zeromq-dev mailing list [email protected] http://lists.zeromq.org/mailman/listinfo/zeromq-dev
