Cogent Communications may be the source of the lag guys, NOT SE. Please research:

[Raist]

Seriously, they are KNOWN offenders. Run a trace to your World Servers and see if cogentco shows up in your route just before it starts hitting Ormuco... if you see cogentco, that is likely a major player in this problem. If it isn't Cogent, but someone else and you see a similar pattern as what shows in the following trace logs, research those IP addresses or FQDN names to find out who is responsible and if they too have been documented for having lossy routing.

Tracing route to 199.91.189.74 over a maximum of 30 hops

<<First few hops removed for security reasons>>

5 31 ms 30 ms 30 ms bu-ether14.atlngamq46w-bcr00.tbone.rr.com [66.109.6.82]
6 29 ms 27 ms 27 ms ae-0-0.pr0.atl20.tbone.rr.com [66.109.6.171]
7 77 ms 78 ms 79 ms te0-0-0-10.ccr21.atl02.atlas.cogentco.com [154.54.12.109]
8 77 ms 80 ms 81 ms be2052.mpd21.atl01.atlas.cogentco.com [154.54.40.249]
9 86 ms 91 ms 90 ms te0-0-0-7.mpd21.dca01.atlas.cogentco.com [154.54.28.197]
10 107 ms 96 ms 93 ms te0-1-0-7.ccr21.jfk02.atlas.cogentco.com [154.54.1.234]
11 122 ms 114 ms 123 ms be2108.ccr21.ymq02.atlas.cogentco.com [154.54.3.134]
12 78 ms 68 ms 69 ms 38.122.42.34
13 68 ms 68 ms 68 ms 10.2.2.1
14 69 ms 69 ms 69 ms 192.34.76.2
15 73 ms 68 ms 78 ms 199.91.189.234
16 68 ms 69 ms 69 ms 199.91.189.74

Trace complete.

Here's a post from a Blizzard thread from about a year ago, titled "Lagspikes and the damn atlas.cogentco.com":

http://us.battle.net/wow/en/forum/topic/6794352287

Location: Australia
ISP: TPG
Problem: Ridiculous lag spikes

Trace route when normal:

Tracing route to 184.254.129.12.in-addr.arpa [12.129.254.184]
over a maximum of 30 hops:

1 <1 ms <1 ms <1 ms 1.1.168.192.in-addr.arpa [192.168.1.1]
2 19 ms 15 ms 14 ms 49.21.20.10.in-addr.arpa [10.20.21.49]
3 20 ms 15 ms 15 ms 220-244-137-61.tpgi.com.au [220.244.137.61]
4 15 ms 15 ms 30 ms syd-nxg-men-crt1-ge-7-1-0.tpgi.com.au [202.7.162.105]
5 187 ms 197 ms 187 ms te8-3.ccr01.sjc05.atlas.cogentco.com [38.122.92.41]
6 184 ms 185 ms 185 ms te0-3-0-2.mpd22.lax01.atlas.cogentco.com [154.54.25.189]
7 185 ms 185 ms 186 ms te4-4.ccr02.lax05.atlas.cogentco.com [154.54.82.162]
8 184 ms 183 ms 196 ms att.lax05.atlas.cogentco.com [154.54.11.10]
9 188 ms 186 ms 186 ms cr1.la2ca.ip.att.net [12.122.84.202]
10 183 ms 185 ms 191 ms gar20.la2ca.ip.att.net [12.122.128.181]
11 203 ms 187 ms 184 ms 12-122-254-238.attens.net [12.122.254.238]
12 184 ms 184 ms 184 ms mdf001c7613r0004-gig-10-1.lax1.attens.net [12.129.193.250]

Trace route when lagging: (see underlined)

1 <1 ms <1 ms <1 ms 1.1.168.192.in-addr.arpa [192.168.1.1]
2 15 ms 14 ms 15 ms 49.21.20.10.in-addr.arpa [10.20.21.49]
3 30 ms 15 ms 15 ms 220-244-137-61.tpgi.com.au [220.244.137.61]
4 19 ms 14 ms 15 ms syd-nxg-men-crt1-ge-7-1-0.tpgi.com.au [202.7.162.105]
5 177 ms 203 ms 207 ms te8-3.ccr01.sjc05.atlas.cogentco.com [38.122.92.41]
6 185 ms 185 ms 185 ms te0-3-0-2.mpd22.lax01.atlas.cogentco.com [154.54.25.189]
7 382 ms 209 ms * te4-4.ccr02.lax05.atlas.cogentco.com [154.54.82.162]
8 189 ms 183 ms 196 ms att.lax05.atlas.cogentco.com [154.54.11.10]
9 187 ms 186 ms 187 ms cr2.la2ca.ip.att.net [12.122.84.218]
10 182 ms 183 ms 186 ms gar29.la2ca.ip.att.net [12.122.129.241]
11 187 ms 184 ms 196 ms 12-122-254-234.attens.net [12.122.254.234]
12 184 ms 184 ms 184 ms mdf001c7613r0004-gig-10-1.lax1.attens.net [12.129.193.250]

Tracing route to 184.254.129.12.in-addr.arpa [12.129.254.184]
over a maximum of 30 hops:

1 1 ms <1 ms <1 ms 1.1.168.192.in-addr.arpa [192.168.1.1]
2 15 ms 15 ms 15 ms 49.21.20.10.in-addr.arpa [10.20.21.49]
3 15 ms 15 ms 15 ms 220-244-137-61.tpgi.com.au [220.244.137.61]
4 23 ms 15 ms 15 ms syd-nxg-men-crt1-ge-7-1-0.tpgi.com.au [202.7.162.105]
5 365 ms 208 ms 207 ms te8-3.ccr01.sjc05.atlas.cogentco.com [38.122.92.41]
6 193 ms 185 ms 184 ms te0-3-0-2.mpd22.lax01.atlas.cogentco.com [154.54.25.189]
7 * 185 ms 185 ms te4-4.ccr02.lax05.atlas.cogentco.com [154.54.82.162]
8 183 ms 183 ms 202 ms att.lax05.atlas.cogentco.com [154.54.11.10]
9 211 ms 186 ms 187 ms cr2.la2ca.ip.att.net [12.122.84.218]
10 277 ms 206 ms 199 ms gar29.la2ca.ip.att.net [12.122.129.241]
11 185 ms 184 ms 184 ms 12-122-254-238.attens.net [12.122.254.238]
12 184 ms 184 ms 184 ms mdf001c7613r0004-gig-10-1.lax1.attens.net [12.129.193.250]

The thing is, this isn't the first time, and it's not rare to happen at all either. It's been happening for years and whenever it gets resolved, same thing happens again in a few months.

There are additional posts in there that help to further emphasize the scope of the problem. This isn't an isolated incident either. If you spend some time googling, you will see it is a common problem. Players getting routed through Cogentco frequently have issues.

Notice the extended response times for JUST the cogentco segments. This is a very typical thing when passing through their segments. If you spend some time researching their routing, you will quickly see a pattern of them being reported for having high packet loss, and it gets progressively worse during peak traffic windows.

Edit:
**It should be noted, that I don't generally have lag issues in game. So, my response times may be considerably low compared to some of you guys having real issues. What you are looking for is the marked change in one particular segment where response time gets really high (possibly with some time outs), then drops really low again. This is where there is potentially a networking issue that needs to be investigated.**

Cogent has about a dozen or so routers in Quebec, and a bunch of them are in/around Montreal. They are frequently reported for having high loss rates--consistently in excess of 25%. There have been numerous complaints posted about them in various forums over the years.

Please... if you guys are serious about getting this resolved for you, run traces to your world servers (can get the IP's via something like netstat or TCPView). Regardless of what they show, if you are experiencing these high lag/connectivity issues, then copy/paste your results into notepad or something and forward the log to BOTH your ISP and SE when you request assistance from them. Our best hope at resolving at least this fixable aspect of the communications problems is to get the right people leaning on the people responsible for such faltering network segments between us and SE.

After THAT gets resolved, if the problem still persists, then we can start leaning on SE for further tweaking, as we will have proven it is NOT an ISP issue (be it our ISP, SE's ISP, or a segment between us) because we will be able to show a quick, responsive route between the endpoints.

I can't stress this enough. Until we can rule out such issues with interconnectivity, everyone will keep pointing fingers at each other (SE and ISP), and NOTHING will be done to resolve these actual networking problems. If we get a little proactive instead of just b*tching about the lag, we might actually start working towards solving this problem for you guys.

[Ramzi]

Mine doesn't run through cogentco, and I'm still getting major lag. My highest ping was 41ms and there were 11 hops total. Final named hop was through mtl.iweb.com

[Raist]

could you post your trace as well? Perhaps we can find another pattern here as well. I just know from past experience that Cogent has issues, so when I saw them in some traces on the forums, I looked them up again and they were showing 50% loss at times a couple weeks ago. Figured I'd put it out there for everyone to investigate on their lines.

There is a common denominator here, as some people are experiencing issues while others on the same server are not--or at least not to the extent that others are seeing. If we can gather/submit enough samples, perhaps a pattern can be discerned and someone can finally take some action towards addressing the problems.

[Haleir]

I know google exists, but perhaps if you throw in some brief instructions with regards to command prompt and how to get one's server's info so one may run traceroute, more people would be willing to add to the data pool.

I'd be helpful and add such instructions myself, but while I know how to do some of it, I don't know how to do all of it. sorry. :P

[Max_Ohm]

As far as how to obtain specific ffxiv server's ip addresses, I can't help you there.

Running a command prompt tracert when you have found the ip address is very easy though.

Simply hit your windows button, and in the field where is says "Search Programs or Files, type cmd
After it brings up the command prompt, type "tracert 180.170.170.88" (no quotes, and that ip number is just an example)

[Haleir]

Thanks for the reply Max!
Alas that is the part I've managed to work out. It's grabbing a FFXIV server address that eludes me, at present.
I'm sure your primer will enlighten some others though.

[Max_Ohm]

This information may be more helpful, however there are still more specific IP addresses assigned to each ffiv server (world):

North America Datacentre | square-enix.montreal.com | IP address : 184.107.107.176
Europe Datacentre | eu.square-enix.com | IP address : 209.130.141.243
Japan Datacentre | jp.square-enix.com | IP address : 202.67.53.202

I would like to add that it is not only Cogent that has this issue although I have had bad hops when trying to move through a cogent server. My most frequent offender seems to be a server in New York (if-1-2.tcore1.nyy-newyork.as6453.net/) though I'm having trouble finding out who it belongs to.

It may also be prudent to keep your prompt open in the background while you are playing. The moment you experience severe lag or a 90000 error, switch to the prompt and run another tracert. I notice that these "bad hops" for me do not always occur so it can be hard to catch it.

[Max_Ohm]

Going to bump this again with some more anecdotal evidence of probable SE server fatigue.

Spent the entirety of the day with continuous lag spikes ( I was in Mor Dhona mostly, probably the most populated area on Behemoth at the moment). Tons of error 90000 all day, with a light sprinkling of Pure Virtual Function Call error. I decided to play a bit on my alt char on a Japanese server (a low population one). Not a single issue with lag or getting disconnected..I've been on for hours without even a hic up. NPCs appear right away..no character lag, no disconnections.

Take from this what you will. The minute I try to play my main on Behemoth, the connection just goes all to hell again.

[Raist]

I know google exists, but perhaps if you throw in some brief instructions with regards to command prompt and how to get one's server's info so one may run traceroute, more people would be willing to add to the data pool.

I'd be helpful and add such instructions myself, but while I know how to do some of it, I don't know how to do all of it. sorry. :P

Sorry.. figured all these guys knew this stuff already. Usually pretty common knowledge amongst PC enthusiasts (or at least the ones predating Vista).

You can easily get the IP's by running Resource Monitor (resmon.exe in run or search box), and clicking the Network tab. FFXIV should be listed with the IP address(es) it is connecting to. The ones starting with 199.91.xxx.xxx will be the world servers and such in Canada. The ones starting with 124.150.xxx.xxx will be servers in Japan that it also maintains connections to (lobby server and such).

If you can't get at it via Resource Monitor (forget how it looked in older Windows), just open the CMD prompt (as already stated, type cmd in the run/search box to get at it). When logged in the game, just run the command netstat and it will spit out a list of the IP's currently active or recently active and in a wait state. You will be looking for the same IP's there.

Once you have your IP's just do as already mentioned in the CMD prompt: tracert 199.91.xxx.xxx.

You can run the cls command to clear all the text if you want to wipe it and do a fresh trace for copying/pasting it to notepad or email or something. I find dragging to highlight text is sometimes frustrating on a laptop's touchpad, so usually clear the text, trace, then select all text for copying--if you don't clear the text, you can wind up with a lot of junk when you do this. You can select all text and copy it from the cmd prompt window by clicking the black/white C:\ icon in the upper-left corner of the window (pressing the <enter> key will copy all highlighted text also, so you don't have to open the menu again once you go to Edit -> Select all text--just hit the <enter> key to put it on clipboard). Then you can just paste it into documents and such as usual (CTRL-V shortcut or use the other editing options to paste text).

[Raist]

Going to bump this again with some more anecdotal evidence of probable SE server fatigue.

Spent the entirety of the day with continuous lag spikes ( I was in Mor Dhona mostly, probably the most populated area on Behemoth at the moment). Tons of error 90000 all day, with a light sprinkling of Pure Virtual Function Call error. I decided to play a bit on my alt char on a Japanese server (a low population one). Not a single issue with lag or getting disconnected..I've been on for hours without even a hic up. NPCs appear right away..no character lag, no disconnections.

Take from this what you will. The minute I try to play my main on Behemoth, the connection just goes all to hell again.

That's just it though... this doesn't necessarily mean your issues are with the servers. You are getting routed to vastly different regions---one is to Canada and Japan, the other is all to Japan. What we have been experiencing appears to be routing isues towards the end of the route, just as it is getting into the region where the servers reside. So... if you are having routing problems geting to both a Canadian region AND Japanese region from say, Australia at the same time---this problem likely will completely disappear when you are being routed strictly to Japan if the problem is in fact tied to specific routing segments only in that Canadian region.

That is what we are talking about trying to isolate. Try to pinpoint just where it is going south for you guys, and where it is NOT for those not afflicted with this horrible lag.

Again, we are seeing players on the same servers that reside in Canada that are not experiencing these issues. In some cases, the two clients are connecting from the same region, with the same ISP--and yet, one has issues and the other does not. Perhaps they are getting routed differently because of flubbed DNS, who knows.

But, that is the information that is needed. Just what are the paths being taken when things go bad, and is there commonality amongst them. If the bad routes are consistently with one provider (or just a select few), then those routers can be targeted and tested for routing and stability issues.