Latency & packet loss during peak hours

I frequently experience high latency and dropped packets while playing, especially during peak hours (although it seems to sporadically happen during less active times as well).

The following tests were all performed during a time when these connectivity issues were occurring:

Performing a trace route to the NA datacenter yielded the following:

Tracing route to 184.107.107.176 over a maximum of 30 hops

1 <1 ms <1 ms <1 ms 192.168.1.1

2 2 ms 2 ms 1 ms L100.TAMPFL-VFTTP-71.verizon-gni.net [173.65.173.1]

3 7 ms 3 ms 2 ms G0-12-1-4.TAMPFL-LCR-21.verizon-gni.net [100.41.206.242]

4 * * * Request timed out.

5 12 ms 12 ms 12 ms 0.ae3.BR1.MIA19.ALTER.NET [140.222.225.111]

6 10 ms 12 ms 12 ms 204.255.168.10

7 81 ms 81 ms 81 ms if-1-2.tcore2.MLN-Miami.as6453.net [63.243.152.62]

8 81 ms 81 ms * if-3-2.tcore2.DT8-Dallas.as6453.net [66.110.72.6]

9 82 ms 84 ms 90 ms if-2-2.tcore1.DT8-Dallas.as6453.net [66.110.56.5]

10 83 ms 105 ms 82 ms 64.86.78.5

11 * 81 ms 81 ms if-22-2.tcore2.CT8-Chicago.as6453.net [64.86.79.1]

12 85 ms 82 ms 80 ms if-3-2.tcore1.W6C-Montreal.as6453.net [66.198.96.45]

13 82 ms 85 ms 83 ms 66.198.96.58

14 81 ms 81 ms 81 ms te7-3.dr9.mtl.iweb.com [67.205.127.214]

15 81 ms 82 ms 81 ms 72.55.128.44

16 82 ms 82 ms 83 ms 184.107.107.176

As hop 4 seems to simply be a router that doesn't respond to ICMP requests, this leaves hops 8 and 11 as our most likely candidates for the issues.

Pinging hop 8 yielded the following:

Ping statistics for 66.110.72.6:

Packets: Sent = 50, Received = 49, Lost = 1 (2% loss),

Approximate round trip times in milli-seconds:

Minimum = 39ms, Maximum = 48ms, Average = 41ms

While a lost packet here is certainly concerning it doesn't quite seem to explain the issues I'm having.

Pinging hop 11 yielded the following:

Ping statistics for 64.86.79.1:

Packets: Sent = 50, Received = 44, Lost = 6 (12% loss),

Approximate round trip times in milli-seconds:

Minimum = 62ms, Maximum = 70ms, Average = 63ms

This seems to be our main culprit. AS6453.net is dropping packets as they pass through Chicago.

I would very much appreciate it if someone from SE could look into this routing issue and do anything they could to encourage packets be routed elsewhere en route from Tampa to Montreal. These connectivity issues are not only completely out of my control, but are having a heavy negative impact on my ability to play and enjoy Final Fantasy XIV.

I'm sure there are other locations where traffic is being routed through here and causing connectivity issues to the NA datacenter.

Thank you for your time.

ISP is biggest part of LAG issue

Here is an export from WinMTR from last night (7/21/2015)... earlier in the day there wasn't an issue but about 5pm PST the LAG fest started and by about 6:30pm PST things were getting unplayable... 20 to 30 sec. of I can move but nothing else on the screen was and we wiped several times trying to do a dungeon...

If you look at the second line in the list below you'll see "L100.LSANCA-VFTTP-89.verizon-gni.net" shows and average latency of 12 ms but spikes as high as 1194 ms and this is Verizons equipment, its what my home router directly connects to. This connection normally runs with less than 50 ms as the highest spike but that changes between 5 and 6 pm PST for me and continues for the next 4 to 5 hrs...

Code:

| WinMTR statistics | | Host - % | Sent | Recv | Best | Avrg | Wrst | Last | |------------------------------------------------|-------|-------|------|------|------|------| | Wireless_Broadband_Router.home - 0 | 12271 | 12271 | 0 | 0 | 6 | 0 | | L100.LSANCA-VFTTP-89.verizon-gni.net - 1 | 12266 | 12265 | 3 | 12 | 1194 | 13 | | G0-3-3-4.LSANCA-LCR-22.verizon-gni.net - 1 | 12251 | 12246 | 7 | 12 | 49 | 12 | |so-3-1-0-0.LAX01-BB-RTR2.verizon-gni.net - 1 | 12251 | 12246 | 7 | 15 | 104 | 9 | | No response from host - 100 | 2455 | 0 | 0 | 0 | 0 | 0 | | 0.ae2.XL4.LAX15.ALTER.NET - 1 | 12254 | 12250 | 7 | 25 | 220 | 9 | | TenGigE0-7-0-1.GW5.LAX15.ALTER.NET - 1 | 12262 | 12260 | 6 | 15 | 38 | 16 | | kddia-gw.customer.alter.net - 1 | 12258 | 12255 | 17 | 24 | 303 | 18 | | lajbb001.int-gw.kddi.ne.jp - 1 | 12262 | 12260 | 16 | 22 | 136 | 18 | | otejbb205.int-gw.kddi.ne.jp - 1 | 12262 | 12260 | 129 | 133 | 232 | 131 | | jc-ote301.int-gw.kddi.ne.jp - 1 | 12262 | 12260 | 117 | 121 | 191 | 120 | | 106.187.28.198 - 1 | 12251 | 12246 | 110 | 114 | 232 | 112 | | 61.195.56.133 - 1 | 12239 | 12231 | 108 | 112 | 254 | 111 | | 219.117.144.78 - 1 | 12255 | 12251 | 121 | 126 | 224 | 125 | | 219.117.144.57 - 1 | 12259 | 12256 | 109 | 111 | 286 | 113 | | 219.117.144.41 - 1 | 12254 | 12250 | 109 | 119 | 176 | 124 | | 219.117.147.194 - 1 | 12262 | 12260 | 113 | 117 | 316 | 119 | | 124.150.157.30 - 1 | 12262 | 12260 | 109 | 113 | 126 | 113 |

184.107.107.176 has nothing to do with the game. Not the same ISP and a different section of town, so your routing is different.

And don't assume the IP at arrstatus is correct either...they have changed several times and some listed there are not in use atm. Use the address your client actually connects to while logged fully into the game (can check with resmon from search/run box) or use one of the lobby server addresses: neolobby01.ffxiv.com through neolobby06.ffxiv.com with odd being Japan and even are Canada. Those may change though once the EU data center comes online.

Just so we are all on the same page I pulled the IP from the resource monitor as suggested and it shows the second hop "L100.LSANCA-VFTTP-89.verizon-gni.net" Verizons equipment with 1228 ms this time :eek: the 2 Montreal hops are a bit high also but I'd think they should be the highest not the first hop to my ISP...

Can you say THROTTLE :mad:

http://i1345.photobucket.com/albums/...psrw2rkveb.jpg

From what I read from others who posted similar issues in the past year or so, it seems Verizon and other ISPs throttles connections during prime time since so many people use that time to watch videos, play games, etc. The pipelines get choked and the ISPs don't do anything about it; apparently Verizon and others were heavily against net neutrality and wanted companies like Hulu and Netflix to pay for a "faster" lane, AKA "we won't throttle you or your customers if you pay us more for a special faster connection".

Until the ISPs stop being cheap or Square uses someone or something else for their servers, we're stuck like this. I finally caved and gotten a VPN just so I don't lag like crazy during prime time hours. I can't afford to lag in more intense fights like the Alexander raids, especially since I main as a healer.

I'm having lag delay issues since last patch, and it is not just during peak hours. I'd find it highly coincidental if my ISP just started deciding to throttle their service, 24/7, at exactly the same time the patch was implemented. The only other time I've had lag delay issues like this on FFXIV was once during another heavy server congestion, until this last patch.

Any suggestions on getting my ping batchfile working so I can paste my pingtest for 64.86.79.1?

Like the OP I am seeing a lot of lost packets there and here: 64.86.78.5

The first belongs to Tata Communications and the second is tcore, whomever they are. The last during peak here is unplayable.

havent been able to play the game since alex came out. lag is so bad it drops to 0 ping every 10 secs. where do i go to get a credited for the month?

The earlier people in the thread seem to claim that the 2nd hop to Verizon's equipment is problematic.

So let me submit my WinMTR that says something different:

https://i.imgur.com/pzXfHFD.png

This was done last night around 10pm while game was lagging (2s delay on teleport). While a spike to 50ms for the Verizon equipment is less than ideal, a 10-16% packetloss for every hop from 6-14 seems worse (every TATA Communications and Ormuco hop).

Quote:

Originally Posted by keLston

The earlier people in the thread seem to claim that the 2nd hop to Verizon's equipment is problematic.

So let me submit my WinMTR that says something different:

https://i.imgur.com/pzXfHFD.png

This was done last night around 10pm while game was lagging (2s delay on teleport). While a spike to 50ms for the Verizon equipment is less than ideal, a 10-16% packetloss for every hop from 6-14 seems worse (every TATA Communications and Ormuco hop).

I get the same packet loss and I am on Level 3 peering. It's FiOS. They don't want to take responsibility but that is who the problem is.

That .57 address is no longer a valid IP. Has not hosted a world service since infrastructure changes back in March/April. Need to get the IP your client actually uses or use the registered neolobby addresses.

Also...your routed differently in general. You are in different areas (NY vs Cali for heavens sake). They do however exhibit similar symptoms at the peer exchanges--even though one is handing off to Level3 and the other is handing off to TATA. Looks like more of the same stuff we've seen in the past.