Pinging Armagoshdarn I

n54 · Post by **n54** » Sun Feb 01, 2004 1:40 pm

part I

there is sometimes a lot of bitching about ping on armagoshdarn so i decided to see what the true ping is for me to get an idea on how accurate the server pings are in general

the armagoshdarn server is located at: armagetron.mars.org
and uses port: 4534

checking ping:

Code: Select all

C:\>ping armagetron.mars.org

Pinging armagetron.mars.org [209.126.142.251] with 32 bytes of data:

Reply from 209.126.142.251: bytes=32 time=187ms TTL=241
Reply from 209.126.142.251: bytes=32 time=171ms TTL=241
Reply from 209.126.142.251: bytes=32 time=172ms TTL=241
Reply from 209.126.142.251: bytes=32 time=172ms TTL=241

Ping statistics for 209.126.142.251:
    Packets: Sent = 4, Received = 4, Lost = 0 (0% loss),
Approximate round trip times in milli-seconds:
    Minimum = 171ms, Maximum =  187ms, Average =  175ms

i've run this some times now at different hours and seem to be getting close to 188ms on average in total. 188ms (milliseconds) is the same as 0.188 seconds, barely a tenth of a second.

i've tested this while getting ping values from the server ingame spanning such different values as 40ms, 287ms and 401 ms. at the same time some people get ping values of 0ms(!), 2ms, 9ms, and so on... now everyone who has been doing some network testing/troubleshooting knows that all machines use at least something like 8ms just to reach the loopback address in the network card (this is hardware dependant and not system (os) dependant), so obvioulsy the pings given by armagetron has to be taken as very approximate

ping flaming is lame, the fools who do it are shitting their pants without realizing it

checking route

Code: Select all

C:\>tracert armagetron.mars.org

Tracing route to armagetron.mars.org [209.126.142.251]
over a maximum of 30 hops:

  1   <10 ms   <10 ms   <10 ms  *local internal address*
  2    16 ms    15 ms    16 ms  129.80-202-164.nextgentel.com 80.202.164.129]
  3    16 ms    15 ms    16 ms  217-13-7-50.dd.nextgentel.com [217.13.7.50]
  4    16 ms    15 ms    16 ms  217-13-7-36.dd.nextgentel.com [217.13.7.36]
  5    16 ms    16 ms    15 ms  217-13-6-166.dd.nextgentel.com [217.13.6.166]
  6    15 ms    16 ms    15 ms  ge0-3-0-103.ar1.OSL1.gblx.net [208.51.41.57]
  7    15 ms    32 ms    15 ms  pos12-0-2488M.cr1.CPH1.gblx.net [67.17.72.238]
  8   125 ms   125 ms   125 ms  pos3-2-2488M.cr2.WDC2.gblx.net [67.17.74.78]
  9   125 ms   125 ms   125 ms  so5-1-0-2488M.ar1.DCA3.gblx.net [67.17.68.37]
 10   125 ms   125 ms   125 ms  208.51.74.6
 11   172 ms   188 ms   187 ms  hrndva1wcx2-pos0-0.wcg.net [64.200.89.49]
 12   172 ms   187 ms   188 ms  anhmca1wcx3-oc48.wcg.net [64.200.240.29]
 13   172 ms   188 ms   187 ms  sndgca1wce2-pos3-0.wcg.net [64.200.141.18]
 14   203 ms   187 ms   188 ms  sndgca1wce1-calfregional-gige.wcg.net [65.77.90.42]
 15   172 ms   187 ms   188 ms  surveyor.mars.org [209.126.142.251]

Trace complete.

i checked the route to see how far away from the server i was in router hops, 15 isn't that bad (= 14 different different stretches of network wire travelled) and only through a maximum of 5 different domains (w/subdomains).

also the tracert reveals that the armagoshdarn server has an alias "surveyor.mars.org", for those of you who have been at the main mars.org site you know that it is a "sleeping" site with only a search box linked to a pseudo-search engine, basically just a site to ensure ownership of domain names (www.mars.org, surveyor.mars.org and probably lots of others as well), it might get a lot of unwarranted traffic from web typos and poor search engines and this might explain some of the lag, at least when the server "is ill".

n54 · Post by **n54** » Sun Feb 01, 2004 1:41 pm

part II

definitions on lag can be found here http://whatis.techtarget.com/definition ... 56,00.html , i'm using the term lag as in perceived lag

well that's all i cared to do for now

Post by **Tank Program** » Sun Feb 01, 2004 2:23 pm

What you maybe forgetting, is that armagetron in a way changes your ping...

CYCLE_PING_RUBBER x niceness when hitting a wall, influence of your ping

Post by **dlh** » Sun Feb 01, 2004 3:40 pm

ping results
let it run while sitting in spectator mode and playing

Code: Select all

--- armagetron.mars.org ping statistics ---
741 packets transmitted, 741 packets received, 0% packet loss
round-trip min/avg/max = 79.158/96.474/365.432 ms

traceroute

Code: Select all

traceroute armagetron.mars.org
traceroute to armagetron.mars.org (209.126.142.251), 30 hops max, 40 byte packets
 1  10.222.160.1 (10.222.160.1)  13.212 ms  21.315 ms  13.558 ms
 2  24.218.190.185 (24.218.190.185)  8.143 ms  7.932 ms  9.279 ms
 3  24.218.190.137 (24.218.190.137)  10.029 ms  24.325 ms  9.937 ms
 4  24.218.190.133 (24.218.190.133)  10.024 ms  29.443 ms  8.969 ms
 5  24.218.190.129 (24.218.190.129)  120.022 ms  11.223 ms  9.926 ms
 6  bar01-p0-0.brtnhe1.ma.attbb.net (65.96.0.249)  9.145 ms  13.019 ms  98.718 ms
 7  bic01-p4-0.ndhmhe1.ma.attbb.net (24.91.0.177)  65.08 ms  13.757 ms  10.468 ms
 8  12.125.33.33 (12.125.33.33)  21.022 ms  39.892 ms  11.658 ms
 9  gbr1-p60.cb1ma.ip.att.net (12.123.40.138)  11.662 ms  42.59 ms  10.382 ms
10  tbr1-p013402.cb1ma.ip.att.net (12.122.5.53)  11.523 ms  30.015 ms  18.566 ms
11  tbr2-cl1.n54ny.ip.att.net (12.122.10.22)  20.505 ms  23.897 ms  37.695 ms
12  ggr1-p370.n54ny.ip.att.net (12.123.1.125)  38.432 ms  22.947 ms  16.276 ms
13  nycmny2wcx2-pos6-3.wcg.net (64.200.68.41)  29.31 ms  59.557 ms  19.155 ms
14  hrndva1wcx2-oc48.wcg.net (64.200.240.46)  81.291 ms  81.509 ms  83.081 ms
15  anhmca1wcx3-oc48.wcg.net (64.200.240.29)  125.69 ms  78.171 ms  79.767 ms
16  sndgca1wce2-pos3-0.wcg.net (64.200.141.18)  81.219 ms  85.878 ms  81.942 ms
17  sndgca1wce1-calfregional-gige.wcg.net (65.77.90.42)  81.716 ms  91.613 ms  80.901 ms
18  surveyor.mars.org (209.126.142.251)  80.127 ms  84.305 ms  133.717 ms

n54 · Post by **n54** » Sun Feb 01, 2004 6:26 pm

no, ping is ping and nothing else (an icmp packet sent to the server, processed, and sent back), if the program uses ping in another sense it isn't ping anymore and shouldn't be termed as ping.

but i don't think the program makes this mistake: afaik the meaning of your quote from the documentation has to do with the way the server figures out cycle_rubber based on what ping value you have rather than changing the ping value itself. ie. you get more cycle_rubber if you're ping is high, this to even out differences between people with low and high ping when it comes to local rubber perception. get what i mean?

what i'm talking about is that:

CYCLE_PING_RUBBER x niceness when hitting a wall, influence of your ping

is a typo/esl mistake and should be read as:

CYCLE_PING_RUBBER x niceness when hitting a wall, influenced by your ping

but the ping values themselves as seen in the ingame score page (tab key by default) are obviously broken and that part of the code should be replaced (may have something to do with the other issues in the netcode?)

btw my ping examples were from both when playing and not

Post by **Tank Program** » Sun Feb 01, 2004 10:07 pm

I have an idea of what maybe, just maybe be causing some of this... But first, thank you for clearing that bit with the ping influenced stuff, that makes it clearer to me. As for why the ping values are distorted in the game, maybe it is because A) the armagetron dedicated server takes a bit of cpu, and B) it's sending out ICMP packets to on average 6 clients, and then they have to get back, maybe it's loosing some along with the rest of the udp traffic, causing ping distortions... But that's just an idea.

iceman · Post by **iceman** » Sun Feb 01, 2004 11:12 pm

i dont think that using the dos ping command is a good test of server lag because i doubt that the data packed send by dos ping is the same length as the data sent to/from armagetron server and im sure theres other factors involved here too

n54 · Post by **n54** » Sun Feb 01, 2004 11:37 pm

i'll repeat myself: ping is ping no matter what, i don't care if you ping from a mobile phone: it's still a special icmp packet, and the server does just the same amount of work with that packet no matter what else... how long the processing of the packet takes depends (of course) on the server workload. but the server workload doesn't change much by pinging once or twice using the command prompt in addition to playing...

it's not like we're talking about the ping of Death here (old type of DoS (Denial of Service) attack), which requires quite a lot of computers to spew out ping requests constantly.

if there was any confusion: i pinged both while playing and not, and iceman i haven't said one word about testing lag, i'm testing ping (i know some ppl think they're the same but they're not, check the link. ping can be (and is) used as an indicator of latency (note i didn't use the word lag), of course, but nothing more (ie. it can't be used as a measurement of perceived lag)

i haven't checked but there might be an RFC on ping, maybe i'll google it one day

iceman · Post by **iceman** » Mon Feb 02, 2004 12:00 am

hey i didnt say that ping and lag were the same thing

i said that using dos ping does NOT indecate server lag

n54 · Post by **n54** » Mon Feb 02, 2004 12:00 am

RFC 792 ICMP page 14 mainly

+ ok iceman double misunderstanding there, just so used to ppl saying ping is lag *grr* sorry about misinterpreting you lol

+ and this one was a nice read http://ftp.arl.mil/~mike/ping.html

iceman · Post by **iceman** » Mon Feb 02, 2004 12:11 am

n54 · Post by **n54** » Mon Feb 02, 2004 12:16 am

iceman · Post by **iceman** » Mon Feb 02, 2004 12:20 am

n/p dude, im shit at explaining stuff so its easy to misunderstand me

hey i confuse myself most of the time ! lol

n54 · Post by **n54** » Mon Feb 02, 2004 1:57 am

don't worry about it, was my fault too

Post by **Tank Program** » Mon Feb 02, 2004 8:59 am

Iceman: Nice with the smilies...

Armagetron Forums

Pinging Armagoshdarn I

Pinging Armagoshdarn I

ping

i didnt say that

jello

hey