<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<div class="moz-cite-prefix">Depends on your nic (or upstream)
buffers, both rx/tx, and how efficient they, or upstream process
opening a socket, can consume data. If the socket (ie. tcp/udp
ports) fill first, this usually amounts to kernel-level
congestion. If the hardware buffer on the card and/or queues in
use fill first, than it's a hardware issue.<br>
<br>
Either way, queues and buffers are tunable usually, no one really
bothers however.<br>
<br>
Longer your bits stays in queue, or eventually drop, you introduce
"lag". Lag is a result of buffers not answering in a timely
fashion, or being forced to drop and signal a retransmit, and this
could be your nic, theirs, or your upstream modem/provider. <br>
<br>
The fact you're saying that you see "everyone" slow down means
probably yours is to blame more than everyone else. In your case
tx-queue buffers are likely overloading as it can't transmit the
data fast enough, or you need to increase your tx buffers at the
kernel level. On linux, that's a sysctl, namely wmem stats under
net:<br>
<br>
mb@host:~# sysctl -a | egrep 'rmem|wmem'<br>
net.core.rmem_default = 212992<br>
net.core.rmem_max = 212992<br>
net.ipv4.tcp_rmem = 4096 87380 6291456<br>
net.ipv4.udp_rmem_min = 4096<br>
net.core.wmem_default = 212992<br>
net.core.wmem_max = 212992<br>
net.ipv4.tcp_wmem = 4096 16384 4194304<br>
net.ipv4.udp_wmem_min = 4096<br>
vm.lowmem_reserve_ratio = 256 256 32<br>
<br>
Also, txqueuelen under your interface as well can affect this.
You'll usually see drops here from exhaustion, but increasing this
is necessary for speeds above 1gb.<br>
<br>
mb@host:~# ifconfig em1<br>
em1 Link encap:Ethernet HWaddr xx:xx:xx:xx:xx:xx<br>
UP BROADCAST MULTICAST MTU:1500 Metric:1<br>
RX packets:0 errors:0 dropped:0 overruns:0 frame:0<br>
TX packets:0 errors:0 dropped:0 overruns:0 carrier:0<br>
collisions:0 <b>txqueuelen:1000 </b><br>
<br>
Windoze does this somewhere between the registry and the driver
stack as a gui for them. <br>
<br>
Lots can, if tcp-based, tcp congestion algo's can behave
differently too, if your app/game uses tcp.<br>
<br>
Plenty of how-to's out there around stack tuning, usually more for
a) weird apps or b) apps that really need that 1g/10g/40g
connection requiring the tuning for more.<br>
<br>
Any way you look at, efficient pipelining of data is necessary, or
you get lag, drops, etc mentioned above. Most consumers, as well
as developers of network code just expect infinite bandwidth and
omniscient handling of random flows. Truth is, most nics suck
having minimal buffers, especially consumer ones and most server
ones too. Apps are terrible about streamlining data for network
use too (some usually let you tune your "bandwidth" for this).<br>
<br>
Bigfoot "gaming" nics do this, introducing a linux soc to do queue
and congestion management to avoid buffer drops between your pc
and the ethernet connection. They do a form of "auto quality of
service", prioritizing certain things over another.<br>
<br>
Queue management, ala QoS (quality of service) is meant for
prioritizing traffic like voice/video, but also any "realtime"
protocols, which most games utilize underneath as udp traffic.
Look up tc (traffic control) under linux, or qos gpo policies on
windoze, dscp/cos, etc.<br>
<br>
-mb<br>
<br>
<br>
On 07/03/2015 05:10 PM, Stephen Partington wrote:<br>
</div>
<blockquote
cite="mid:CACS_G9ymThfsoUC7VByVP92D8+hXO3W3BNnkex4GiNw8CMW_wQ@mail.gmail.com"
type="cite">
<p dir="ltr">Mostly it's in how they are making cheaper gigabit
ethernet. </p>
<div class="gmail_quote">On Jul 3, 2015 4:40 PM, "Wayne D" <<a
moz-do-not-send="true" href="mailto:waydavis@centurylink.net">waydavis@centurylink.net</a>>
wrote:<br type="attribution">
<blockquote class="gmail_quote" style="margin:0 0 0
.8ex;border-left:1px #ccc solid;padding-left:1ex">I have a
question regarding ping times in game servers:<br>
<br>
<br>
I've noticed when playing several online games that ping times
go up as more players connect.<br>
I also run a small game server here at my house and have
witnessed this issue first-hand. The processing overhead of
the game is actually quite small even with 12 players and I
suspect it would not go above 25 to 30%.<br>
<br>
This issue doesn't make any sense to me when we are talking
about no more than 12 players connected.<br>
My bandwidth both up and down are sufficient to handle the
traffic with plenty of room to spare.<br>
<br>
The extra overhead on a single pipe in my mind should not make
any difference considering the processor speeds are so much
faster than the data travelling across the Internet. From my
perspective, I should be able to connect 100 different players
and see no change in lag/ping TO EACH PLAYER. I understand
that all players have to go through the protocol stack on the
net card but that shouldn't matter correct?<br>
<br>
Or, is the actual problem the network card itself? Is that the
reason for the lag? If so, is there a type of card that I
should be looking for “high throughput"?<br>
<br>
Could somebody elaborate on the mechanics of this?<br>
<br>
Thanks in advance.<br>
---------------------------------------------------<br>
PLUG-discuss mailing list - <a moz-do-not-send="true"
href="mailto:PLUG-discuss@lists.phxlinux.org"
target="_blank">PLUG-discuss@lists.phxlinux.org</a><br>
To subscribe, unsubscribe, or to change your mail settings:<br>
<a moz-do-not-send="true"
href="http://lists.phxlinux.org/mailman/listinfo/plug-discuss"
rel="noreferrer" target="_blank">http://lists.phxlinux.org/mailman/listinfo/plug-discuss</a><br>
</blockquote>
</div>
<br>
<fieldset class="mimeAttachmentHeader"></fieldset>
<br>
<pre wrap="">---------------------------------------------------
PLUG-discuss mailing list - <a class="moz-txt-link-abbreviated" href="mailto:PLUG-discuss@lists.phxlinux.org">PLUG-discuss@lists.phxlinux.org</a>
To subscribe, unsubscribe, or to change your mail settings:
<a class="moz-txt-link-freetext" href="http://lists.phxlinux.org/mailman/listinfo/plug-discuss">http://lists.phxlinux.org/mailman/listinfo/plug-discuss</a></pre>
</blockquote>
<br>
</body>
</html>