While there could be issues in your local area, it more likely is an issue with the routing partners you have been assigned to get you across the border into Canada and over to SE's ISP. They also have to use these third parties to connect to other companies within the US as well (TWC customers don't typically interface directly with AT&T--they will go through someone like Cogent to connect the two networks). When they can't arrange a direct peering agreement with a content provider like Netflix, the last mile providers at both ends have peering agreements with a small handful of Tier 1 providers for this transit. For years, our ISP's have neglected to update these peering arrangements to compensate for the increasing traffic flow. Ultimately, it is our ISP's that must address this issue.
You can sort of see the pattern when you look at the outage maps for some of these companies. I will provide some sample maps below, with links to the source. The big 3 Tier1 providers we are getting assigned for this transit (Level3, Tata, Cogent) have the bulk of their access to Canada centered in the North East. Only a few are out west above Washington state...the rest are right there around the NY/NJ and MA areas (also Nova Scotia for the EU). There is also an issue with a NSA scan that happens along the way, that shoots us over towards Chicago--for many this means cutting across/around Ohio/Indiana that is also a hotbed for congestion.
Basically, all of this boils down to outdated route analytics and poorly managed interconnects between the companies involved in getting us from the US and into/across Canada. SE doesn't really have much of any influence here, other than pressing their last mile ISP (Ormuco) to address any identified peering issues in Canada. Unfortunately--the bulk of the problems are being detected on the US, AU, or EU sides of these routes. That puts the ball more squarely in our ISP's court to address. But, you need to arm yourself with good data and contact people further up the chain then the guys that answer when you call in. Get some tracerts/traceroutes to your actual game server (can get the IP by running resmon from run/search box, expanding the TCP/Network sections and look for the address starting with 199 that is being used by FFXIV when fully logged into the game). Make sure to grab samples when things are good and bad if you can. Then you need to push to get through to Tier3 support to see if you can get those reports to them. You may be able to do this through their support forums more effectively.
Yes, I know it seems tedious and quite a bother...but unless we start calling out our ISP's "en masse", this may never change. Both TATA and Level3 have filed formal complaints with regulators over the years trying to address this under the fair practices language, but have not had much success. We have the power of the purse though (so to speak). We pay our ISP's for service, and with that comes a responsibility for them to provide some modicum of an expected quality for that service.
Speed is not quality by the way... LATENCY is quality. Latency is a measure of response time, not transfer speed. Each bit actually physically travels at basically the same speed across the wires--you get more speed because you are able to send larger clusters of those bits. A fatter pipe lets more bits through per cycle. It all actually travels at the same speed across the same materials. Think of it as the difference between a 4-lane and an 8-lane highway with both of them having the same speed limit. One has the potential to get more cars through in the same time simply because it has twice as many lanes. Doesn't matter if you can download across a fat pipe if it is constantly pausing, dropping packets, and having to retransmit portions of a file. In the highway comparison, think of what happens when a major accident clogs a lane, a bad storm causes everyone to slow down, or something like a bridge goes out and everyone has to take a detour through a small rural area--heavy congestion and delays develop, causing lots of problems with people getting to places on time.
That is what is typically happening here. The game runs just as well on a 5Mbit connection as a 100Mbit provided their latency is the same---it simply doesn't transfer enough data to take advantage of all that bandwidth. What is at issue is that the routes are congested to the point they are delaying or dropping packets, which results in a lot of retransmits and in extreme conditions even the closing of a session (which can trigger the disconnect).
All of this can be remedied in some fashion by our ISP's. Either they fix their peering arrangements to address the congestion problem--or they switch you to an alternate partner (this can be done.. TWC has been changing me up once or twice a month since I started working with them on this very issue). Ormuco peers with all three of the previously mentioned Tier1's, as well as others. So do the major last mile providers in the US (AT&T, Verizon, TWC, Comcast, etc.). They have options... but you need to get Tier3 involved and show them the proper data (good and bad tracerts) to get them involved. Else, you will keep getting the Tier1/Tier2 techs telling you your local connection is "fine". That's because it likely is... the problem isn't local, but upstream at or beyond the peering level that those guys have nothing to do with. That happens at Tier3 or higher.
And now for some maps. See if you can spot the patterns across them. These images are snapshots of the maps that are updated throughout the day, so they may update periodically when this thread is reloaded. Maps, tweets, support links, and such for various other services (it's not just ISP's) can be found at the downdetector site's companies page:
http://downdetector.com/companies
Comcast Map
http://downdetector.com/status/comcast-xfinity
Verizon Map
http://downdetector.com/status/verizon-communications
AT&T Map
http://downdetector.com/status/att
TWC Map
http://downdetector.com/status/time-warner-cable
Playstation Network Map
http://downdetector.com/status/playstation-network
XBox Live Map
http://downdetector.com/status/xbox-live
And here is a sample of my current route (taken over wifi, so the numbers are a bit skewed versus the hardwired line)--which has been dramatically improved since I opened an investigation with TWC almost 2 weeks ago when they changed me off of Cogentco because it started getting flaky. When I opened this ticket, I had 3 hops having issues, one within the realm of TWC, the others around the handoff into Level3 (ones starting with 4.69). I still have issues with TWC's gateways (start with 24 or 66) that want to get fussy at times, but the real signs of congestion crop up at the Level3 exchanges.
Code:
C:\Windows\system32>tracert -d 199.91.189.25
Tracing route to 199.91.189.25 over a maximum of 30 hops
1 6 ms 1 ms 1 ms 10.10.100.1
2 44 ms 26 ms 27 ms 66.26.112.1
3 26 ms 19 ms 21 ms 24.31.198.9
4 33 ms 25 ms 27 ms 24.93.64.134
5 37 ms 29 ms 30 ms 107.14.19.48
6 32 ms 36 ms 27 ms 107.14.19.99
7 33 ms 26 ms 27 ms 66.109.9.70
8 * 34 ms 29 ms 4.69.150.169
9 67 ms 66 ms 63 ms 4.69.202.242
10 69 ms 65 ms 64 ms 4.69.148.106
11 67 ms 68 ms 65 ms 4.69.143.214
12 70 ms 62 ms 64 ms 4.69.148.49
13 68 ms 66 ms 64 ms 4.69.134.70
14 79 ms 66 ms 64 ms 4.69.148.37
15 88 ms 97 ms 81 ms 4.69.141.5
16 70 ms 64 ms 66 ms 4.69.141.1
17 70 ms 64 ms 64 ms 4.59.178.74
18 55 ms 48 ms 51 ms 10.2.2.1
19 52 ms 59 ms 49 ms 192.34.76.2
20 54 ms 50 ms 51 ms 199.91.189.234
21 62 ms 49 ms 49 ms 199.91.189.25
Trace complete.
(0)
Reply With Quote
