First TRONIC Bowl

[sinewav]

Servers that are West of the Mississippi are a very different animal internet wise than servers east of the Mississippi.

Back that up with some facts so you don't sound like you're talking out of your butt.

All I know is lately TR has been a lagfest. I'm not the only one who's noticed it, and I find their servers are less active than before. I understand there are many factors involved in such a thing - I was simply stating my experience. I'm sure if I wasn't too lazy to do a trace-route I might find the lag that makes it hard for me to play in either TR or Ww.

[owned]

After looking at the board, I realized something. Shouldn't the Ct vs Sp play in a Euro server because we are both euro clans? (correct me sp if I'm wrong) It could be switched with the two us clans (again correct me if I'm wrong) in pru vs. vcl's match.

ok i correct you. pru is european

oops

[compguygene]

ok sine....here are some traceroutes:LEGAL DISCLAIMER: Traceroutes can be VERY misleading, and only tell part of the story. To fully understand how two servers interact, you have to know who their networking providers are! However, I do know who the networking providers are for each of the servers that I am posting a traceroute from and understand the problems, etc. So, I am taking the time to write a book, for not only your sake, but in the hopes that the community will chill! Please understand, I am constantly re-evaluting my VPS providers, and changing VPS providers as needed. For instance, I have just picked up a NYC based VPS that is hosting this server: TheServerPharmNY: Fortress Freedom! I am trying this service out to see if i can improve upon the performance that i currently get with my Chicago VPS. I am also considering dropping my Atlanta VPS and moving what I have there to NYC.

OK, here is the structure of the series of test that I am showing you in this post. I have a VPS that is hosted in the UK. This VPS has provem itself to be a great performer. I will be marketing server slots there. I will run a traceroute from the VPS on Wild West's dedicated server. I will run a traceroute from my Chicago server. I will run a traceroute from my NYC server. Please understand that such traceroutes have are being run between data centers. So, they eliminate the "last mile" that would get to a person's house. The benefit of such a traceroute combined with a knowledge of the networking providers, is thus. It exposes the "underlying structure" of how a given server may connect to someone in said city. It is particularly useful in this scenario because each of these datacenters is very close to an AT&T routing center. AT&T routing centers process 97% of all internet traffic in the USA.

The 1st traceroute is from my Chicago server to my UK server.

compguygene@thserver:~$ traceroute 115.124.117.180
traceroute to 115.124.117.180 (115.124.117.180), 30 hops max, 40 byte packets
1 . (66.90.101.41) 0.037 ms 0.011 ms 0.010 ms
2 75.149.229.81 (75.149.229.81) 1.088 ms 1.064 ms 1.045 ms
3 pos-1-3-0-0-cr01.chicago.il.ibone.comcast.net (68.86.86.33) 1.967 ms 1.950 ms 1.932 ms
4 xe-2-1-0.chi10.ip.tiscali.net (77.67.68.37) 0.927 ms 0.897 ms 0.869 ms
5 xe-10-2-0.lon20.ip.tiscali.net (89.149.185.210) 111.744 ms 111.725 ms xe-10-3-0.lon20.ip.tiscali.net (89.149.184.45) 111.677 ms
6 blueconnex-gw.ip.tiscali.net (77.67.67.138) 111.681 ms 111.612 ms 112.576 ms
7 pri.the-hex.as29550.net (92.48.95.10) 111.544 ms 111.845 ms 111.755 ms
8 pri.bsq3-the.as29550.net (217.112.81.90) 116.719 ms 121.695 ms 121.684 ms
9 94-76-201-90.static.as29550.net (94.76.201.90) 112.678 ms 112.658 ms 111.634 ms
10 115.124.117.180 (115.124.117.180) 112.610 ms 112.240 ms 113.192 ms

People who live in the UK typically ping in the 125ms to 135ms range when using my Chicago server. The company that hosts both servers in this traceroute use the same internet networking providers. And the networking providers are the most commonly used providers by ISP's.

The next traceroute is from my Ww VPS in Texas.

compguygene@compguygene:~$ traceroute 115.124.117.180
traceroute to 115.124.117.180 (115.124.117.180), 30 hops max, 40 byte packets
1 66.71.249.194 (66.71.249.194) 0.055 ms 0.012 ms 0.013 ms
2 core-a-g5-7.atl.wiresix.net (66.71.255.13) 0.815 ms 0.809 ms 0.809 ms
3 wiresix-atl-core-h-p0-1.gnax.net (207.210.92.133) 0.770 ms 0.761 ms 0.748 ms
4 xe1-2.cr01.atl01.mzima.net (67.199.136.157) 13.466 ms 13.462 ms 13.449 ms
5 te1-7.cr01.iad01.us.mzima.net (69.174.120.54) 15.038 ms 15.025 ms 15.012 ms
6 te4-4.cr01.lga01.us.mzima.net (69.174.120.61) 28.716 ms 19.471 ms 19.446 ms
7 xe0-0.cr01.lga02.us.mzima.net (216.193.255.214) 18.934 ms 27.388 ms 27.360 ms
8 eos3-1.cr01.lhr01.uk.mzima.net (216.193.255.178) 98.881 ms eos2-0.cr01.lhr01.uk.mzima.net (216.193.255.46) 98.871 ms 98.858 ms
9 the-linx.blueconnex.net (195.66.224.223) 94.836 ms 94.814 ms 94.777 ms
10 pri.bsq3-the.as29550.net (217.112.81.90) 96.163 ms 96.493 ms 96.135 ms
11 94-76-201-90.static.as29550.net (94.76.201.90) 95.456 ms 95.447 ms 95.521 ms
12 115.124.117.180 (115.124.117.180) 95.512 ms 95.485 ms 95.953 ms

Two things should be noted on this traceroute. First, it is very misleading. Because Wild West's dedicated server uses a very good backbone...it is particularly good at server to server routes. But note network hops 4-8. That mzima.net is a common provider for "West of the Mississippi". However, the company in question is rarely a provider for East Coast or Euro companies. And none of my other VPS providers use this company for an Network provider. The reason that "West of the Mississippi" companies use this company is that they are one of the absolute best providers in that part of the USA. Manta has informed me that his provider is in negotiations with an additional provider that would link up better with the rest of the world. And said provider is used by each of my other VPS providers.

The third traceroute is from my NYC VPS to my UK VPS.

compguygene@TheServerPharmNYarma:~$ traceroute 115.124.117.180
traceroute to 115.124.117.180 (115.124.117.180), 30 hops max, 40 byte packets
1 216.66.73.65 (216.66.73.65) 0.231 ms 0.258 ms 0.320 ms
2 10gigabitethernet2-4.core1.nyc4.he.net (72.52.92.101) 22.070 ms 22.125 ms 22.181 ms
3 10gigabitethernet1-2.core1.lon1.he.net (72.52.92.78) 91.096 ms 91.133 ms 91.190 ms
4 the-linx.blueconnex.net (195.66.224.223) 91.259 ms 91.297 ms 91.340 ms
5 pri.bsq3-the.as29550.net (217.112.81.90) 94.445 ms 94.550 ms 94.677 ms
6 94-76-201-90.static.as29550.net (94.76.201.90) 92.038 ms 91.997 ms 92.031 ms
7 115.124.117.180 (115.124.117.180) 92.104 ms 92.097 ms 92.085 ms

Now you see why I am so hot to get a VPS in NYC! Such a VPS may be ideal for Euro-American interaction. If the server I have put up load tests well. I will list that server for the July Ladle. Initial tests last night with Hoop and a few other Euros have shown that it may be a good performer. But much more gameplay needs to happen before I will entrust my servers to it.

At this point i would like to point out the page on the wiki that refers to lag http://wiki.armagetronad.net/index.php?title=Lag I would like to point out one more issue that the wiki page does not address. During peak usage times of internet use, all bets are off as to the performance of a given game server. It's because of your ISP, not the server. For example, I switched from DSL to cable, because the phone company was rather slow upgrading to a decent sized network pipe. However, you must understand the fundamental advantage of DSL to cable. DSL is a GUARANTEED connection between you and the phone company. You do not share that connection with anyone else. Cable is a shared service of however many households are on the same fiber-optic port at the fiber optic switch your connected to. And many Saturday nights, as more people in my town have gotten Cable Internet, my access sucks. Recently, the phone company upgraded the DSL, so I am going to switch back.

I hope this posts exposes in an understandable manner the underlying networking issues.

[epsy]

$ traceroute 216.66.76.95
traceroute to 216.66.76.95 (216.66.76.95), 30 hops max, 40 byte packets
1 192.168.0.254 (192.168.0.254) 0.426 ms 0.819 ms 0.797 ms
2 78.192.35.254 (78.192.35.254) 1.447 ms 1.612 ms 0.828 ms
3 78.255.5.126 (78.255.5.126) 1.023 ms 1.422 ms 1.059 ms
4 bzn-crs16-1-be1504.intf.routers.proxad.net (212.27.59.21) 2.100 ms 1.657 ms 1.652 ms
5 londres-6k-1-po101.intf.routers.proxad.net (212.27.51.186) 8.626 ms * 8.686 ms
6 amsterdam-6k-1-po100.intf.routers.proxad.net (212.27.56.42) 17.307 ms 17.123 ms 16.367 ms
7 ams-ix.he.net (195.69.145.150) 17.305 ms 17.035 ms 17.200 ms
8 10gigabitethernet4-1.core1.nyc4.he.net (216.66.24.153) 98.933 ms 96.186 ms 96.214 ms
9 10gigabitethernet1-2.core1.chi1.he.net (72.52.92.102) 118.250 ms 121.840 ms 127.618 ms
10 216.66.76.95 (216.66.76.95) 118.657 ms 118.013 ms 118.056 ms

Sadly, it hops through amsterdam, and either hops back and forth to chicago, or actually is in chicago, judging from the hostmask.

[Tank Program]

I understand your reasoning, but I think you're forgetting something. There's a reason it's called the world wide web. It feels like you're assuming fixed routes, that the data will always follow a certain path. This is not true. There are multiple ways for data to get from A to B, these vary depending on traffic, network maintenance, and the gods of the internet. (Also, on who's lately dragged an anchor across an undersea cable.) Let us consider your VPS in NYC. To get from where I live in England, I follow the trail of routers of VirginMedia until I hit upon their uplink to Telia's London data center, where I get routed to their data center in Asheville. From there I go through Hurricane Electric to your NYC data center's uplink to Hurricane Electric.

To get there from my home in Sweden, since it's the home of Telia, I go almost directly from the house (which is on Telia DSL) to their data center in Copenhagen. That goes straight to New York, where there's a cross link from the Telia data center to the Hurricane Electric uplink at your NYC data center.

From a university in Sweden, I go to a data center in Denmark that has an uplink to Hurricane Electric that appears to go directly to your NYC data center.

I've crossed the Atlantic in 3 different ways. If any one route were to fail (damn those ancors) the data would still easily arrive. There's fiber in place that by looking at other tracerotues starts to show how it interlinks. (I was actually very surprised to see the traceroute show I was using Telia from here in England!)

Going to Texas is even more interesting. From here in England I go again via Telia from London to Asheville, but then continue by Telia to Atlanta, hop onto Wiresix, and then to your VPS.

From home in Sweden I again Copenhagen to New York, then to Telia in Asheville and Atlanta, and then Wiresix. From the university I cross over in Sweden to Telia, then the same route as home in Sweden.

Actually at this point, I've had so much fun looking at traceroutes, I'm not sure what my point was anymore. It does seem odd that the IP of your VPS in Texas (66.71.249.194) upon closer inspection appears to be owned by a company, and located in Atlanta.

From your posted traceroute, it seems your Ww VPS routes from Atlanta to Washington DC to New York. Going from the forum's shell in LA goes to Dallas, then Atlanta via Comcast, then cross connects via Mzima to Wiresix.

The reason why I think New York might be a better location for a server is because it's one of the places that cross continental traffic jumps to directly. Washington DC appears to be another one of those locations from playing around with this. (Nearly an hour, eep! This can be way too much fun.) Anywho, I forget my point, but I had to jump in on this. I love discovering network topology. It's what kept me busy all first year at University here.

Note: I'm probably not being strictly true with my usage of data center, but it's easier to think that way.

[compguygene]

Tank, I am aware of the issues that you presented. I merely wanted to give an explanation that would be as understandable to the community at large as possible. And yes, that is why i wanted to experiment with a NYC location. I think it may be ideal for servers that I have been hosting that tend to attract a fairly even mix of Euro and American players. I appreciate the post you made. It said the things I really wanted to say, without me writing another book. As I stated before, only live testing will prove suitability or unsuitability for any given application.

[Word]

all team captains whose Global ID could be found in the wiki article have mod access level on our server now, even after the bowl/ladle unless they abuse it.

[DDMJ]

All team captains have the Team Leader access level in our server with access to these additional commands:

silence
voice

start_new_match

team_name_1
team_name_2

[DDMJ]

Just for fun:

If anyone wants to join my TeamSpeak channel I'm opening it up to the public. 32 people can join and chat away during the Bowl.

Server Address: tronicbowl.ath.cx:8767

[Z-Man]

Recordings:
Quarterfinal DS vs PRU (DS won 2-0)
Semifinal DS vs TR (TR won)

Final TR vs Arrow in two parts. My server borked after one match.
final-tr-arr.aarec
final-tr-arr-2.aarec

As usual, done with 0.2.8.2.1.

[Lackadaisical]

Congrats trix! except for one particularly embarrassing match i had some good games.

[2020]

grats to the winners of the first tronic bowl
twisted rats
or xits or something...
looks like you had an easy time till the final

[Your_mom]

grats to the winners of ladle 22
twisted rats
or xits or something...
looks like you had an easy time till the final

So is ladle 23 going to be called bowl 2? <silly ladle renaming rant not written but implied>

Grats to arrow who played against several tough teams to get to the final. Grats to tr|x| for winning.

[2020]

renaming was for several reasons i guess
primarily to encourage us to increase teams to hit the 32 mark...
and perhaps to have a once every three month biggie...

originally
the tronic progression was purely a numerical increment
spoon with however many
ladle with 100
bowl with 1k players
and cup with 1m players
and we did get over 100 players for this bowl...

and it was rain who revised it in terms of time
i believe he thought of weekly spoons
monthly ladles
seasonal bowls
and annual cup

so
do we go back to ladles
or do we assume we are working with bowls from now on...?
till we get over 1,000 players?
(128 teams!
which would be played over three sundays
if played with one opening round and two blocks of three games each sunday)

[Olive]

Gz Twixted.

couldnt participate in finals though. Sleepy time for me =(.

By the way, playing CT was my best match yet