[ENet-discuss] Reducing traffic from ENet internally on mobile devices

[Krasimir Marinov]

I’d like to join the discussion that had been started 3-4 months ago, because I experience the behaviour described by Stefan.

Monitoring an idle ENet connection with Wireshark reveals that each 500ms a 50B packet (real data is 8B - 4B length = 4B = ENetProtocolPing) is sent from the initiator of the connection
to the “server”. The “server” immediately responds with PING (50B UDP packet, data = 8B).
Immediately after that another packet is sent to the “server”with length 52B (real data is 10B, which makes 10B - 4B length = 6B = ENetProtocolSendReliable).
The server responds immediately with another ENetProtocolSendReliable.

This sequence of packets is repeated every 500ms. I understand the PING, but why is the other packet sent?

P.S. The scheme looks like:

Client                                                                                                                          Server
————————————>50B (ENetProtocolPing)

                                          50B(ENetProtocolPing)<———————————————

———————————>52B(ENetProtocolSendReliable)

                                          52B(ENetProtocolSendReliable)<—————————————



500ms

the same repeats
                                        




Here is the initial email:
Hi Lee and thanks for your response!

But the ping functions get called (after modifying 
ENET_PEER_PING_INTERVAL) every ~30 seconds yet 
enet_peer_queue_incoming_command() is called every second or so on both 
sides. Wireshark shows packets of 50 and 52 bytes size being sent back 
and forth at the same rate.
I'll check what types of commands those are.

Sincerely,
Stefan

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.cubik.org/pipermail/enet-discuss/attachments/20131217/a71827a3/attachment.html>