[zeromq-dev] ZeroMQ behavior in satellite networks (high variable latency)

[Sylvain Fabre - INPIXAL]

Le jeudi 13 f?vrier 2020, 09:20:56 CET Benjamin Henrion a ?crit :
> What kind of latency do you have?
> 
> RTSP is UDP right?
> 
> Or do you use TCP for control messages and UDP for data (I have been
> there before with VoIP with SIP signalling in TCP+SSL and media with
> UDP)?
> 

Hi, 

Latency is around 700ms, with variations between 500 and 900ms.

Yes, RTSP is TCP/UDP per default, but in our use case, we use RTP over TCP, so this the TCP/TCP flavor of RTSP. UDP has proven to be very problematic over this satellite link, unfortunately.

Meanwhile, we discovered that the PULL host (ie the server receiving the packets, Windows 10 per customer request..)  restricts the socket receive buffer, for unknown reasons up to now.
We need to fix this before going further, as it might be the root cause of our issue :) !

Anyway, thanks for your help and hints, 

Sylvain
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20200213/905dbe59/attachment.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 488 bytes
Desc: This is a digitally signed message part.
URL: <https://lists.zeromq.org/pipermail/zeromq-dev/attachments/20200213/905dbe59/attachment.sig>