Request for information: ISPs and P2P throttling

[Glubit]

Name of ISP - Telstra Bigpond
Character name: Jin Marlfox
Country of Residence - Australia
When does the problem occur (always, only during specific times?) - Always
Reference number from your ISP, if you raised this matter with them - Was not given one.

[Mistbane55]

ISP - Telmex
Country - Mexico
When does the problem occur - Everyday after 5p.m.

I ran a traceroute and several pings (getting a 20% loss on peak hours) and i found that the packets get lost on the Tata Comunications segment


1 79 ms 99 ms 98 ms dsldevice.lan [192.168.1.254]
2 60 ms 12 ms 12 ms dsl-servicio-l200.uninet.net.mx [200.38.193.226]

3 47 ms 47 ms 48 ms bb-dallas-bryan-10-tge0-14-0-9.uninet.net.mx [20
1.125.112.158]
4 58 ms * 57 ms ix-7-3-0-0.tcore2.DT8-Dallas.as6453.net [206.82.
142.25]
5 * * * Tiempo de espera agotado para esta solicitud.
6 146 ms * 154 ms if-3-2.tcore1.PDI-PaloAlto.as6453.net [66.198.12
7.25]
7 162 ms 145 ms 145 ms if-1-2.tcore1.NYY-NewYork.as6453.net [66.198.127
.6]
8 146 ms 145 ms 146 ms if-11-2.tcore2.NYY-NewYork.as6453.net [216.6.99.
1]
9 146 ms 145 ms 145 ms if-2-2.tcore2.MTT-Montreal.as6453.net [64.86.226
.13]
10 179 ms 146 ms 145 ms if-0-2.tcore1.MTT-Montreal.as6453.net [216.6.115
.89]
11 145 ms 145 ms 144 ms if-5-2.tcore1.W6C-Montreal.as6453.net [64.86.31.
6]
12 144 ms 144 ms 146 ms 66.198.96.50
13 * 148 ms 145 ms 192.34.76.2
14 * 145 ms * 199.91.189.234
15 144 ms 145 ms 144 ms 199.91.189.74

[KMarsara]

It's hard to say where their instance servers are actually located. I've done some snooping around and I have found some information. Looking over the data I was able to acquire, it looks like Square rents its server space. There seems to be massive latency originating on the Seattle, Chicago, and Montreal (end point) hops owned by as6453.net, which are owned by Tata Communications. Their switches are showing quite of bit of slow down, (220ms+ latency) which in turn is causing more slow down and bottlenecking at the Montreal hops.

Tata Communications are a T1 telecomm in India, but they are at best a T3 in the US. I can't seem to find what T1 is hosting them in the US. If we can find who the backbone they are running through (AT&T, Century Link, Cogent, Level3, NTT, Savvis, SBC, Sprint, Verizon, XO) we can either field our own trouble tickets with them or we can contact our ISP's and request they avoid routing through their hubs.

If servers are being rented though Tata, then we can't reroute around them. We essentially have to deal with the latency until they redistribute their I/O paths or upgrade their switches.

Alternately, it could also be a problem with their MPLS circuits (hosted through Ormuco) reaching maximum bandwidth and causing latency.

The root issue more than likely lies between the two making it difficult to field any fix in a timely manner.

In short, it looks like Square doesn’t own the servers that are housing the North American and European worlds. This combined with the amount of traffic that their host is seeing is causing the latency issues.

3 184.99.64.201 (184.99.64.201) 18.925 ms 17.211 ms 17.251 ms
4 boid-agw1.inet.qwest.net (184.99.65.65) 23.503 ms 17.396 ms 36.997 ms
5 sea-brdr-02.inet.qwest.net (67.14.41.18) 44.756 ms 31.634 ms 37.997 ms
6 ix-1-0-0-0.tcore1.00s-seattle.as6453.net (64.86.123.77) 52.807 ms 304.980 ms 293.516 ms
7 if-0-0-0-2.core1.00s-seattle.as6453.net (64.86.123.2) 463.000 ms 294.448 ms 256.501 ms
8 if-9-3-3-0.tcore2.ct8-chicago.as6453.net (64.86.124.34) 323.445 ms 142.922 ms 96.168 ms
9 if-3-2.tcore1.w6c-montreal.as6453.net (66.198.96.45) 95.991 ms 96.259 ms 95.309 ms
10 66.198.96.50 (66.198.96.50) 99.636 ms 99.195 ms 99.700 ms
11 192.34.76.2 (192.34.76.2) 99.474 ms 98.574 ms 99.333 ms

whois 199.91.189.38
Ormuco ORMUCO-FIBER-NETWORK (NET-199-91-184-0-1) 199.91.184.0 - 199.91.191.255
Eidos ORMUC-SE (NET-199-91-189-0-1) 199.91.189.0 - 199.91.189.255

64 bytes from 199.91.189.38: icmp_seq=290 ttl=243 time=101.808 ms
64 bytes from 199.91.189.38: icmp_seq=291 ttl=243 time=101.363 ms
64 bytes from 199.91.189.38: icmp_seq=292 ttl=243 time=320.852 ms
64 bytes from 199.91.189.38: icmp_seq=293 ttl=243 time=422.690 ms
64 bytes from 199.91.189.38: icmp_seq=294 ttl=243 time=422.983 ms
64 bytes from 199.91.189.38: icmp_seq=295 ttl=243 time=484.630 ms
64 bytes from 199.91.189.38: icmp_seq=296 ttl=243 time=424.351 ms
64 bytes from 199.91.189.38: icmp_seq=297 ttl=243 time=445.440 ms
64 bytes from 199.91.189.38: icmp_seq=298 ttl=243 time=101.748 ms
64 bytes from 199.91.189.38: icmp_seq=299 ttl=243 time=101.246 ms

64 bytes from 199.91.189.38: icmp_seq=15 ttl=243 time=101.559 ms
64 bytes from 199.91.189.38: icmp_seq=16 ttl=243 time=100.622 ms
64 bytes from 199.91.189.38: icmp_seq=17 ttl=243 time=451.749 ms
64 bytes from 199.91.189.38: icmp_seq=18 ttl=243 time=104.527 ms
64 bytes from 199.91.189.38: icmp_seq=19 ttl=243 time=101.387 ms

64 bytes from 199.91.189.38: icmp_seq=25 ttl=243 time=101.613 ms
64 bytes from 199.91.189.38: icmp_seq=26 ttl=243 time=102.453 ms
64 bytes from 199.91.189.38: icmp_seq=27 ttl=243 time=402.926 ms
64 bytes from 199.91.189.38: icmp_seq=28 ttl=243 time=102.158 ms
64 bytes from 199.91.189.38: icmp_seq=29 ttl=243 time=101.519 ms

64 bytes from 199.91.189.38: icmp_seq=60 ttl=243 time=103.801 ms
64 bytes from 199.91.189.38: icmp_seq=61 ttl=243 time=100.424 ms
64 bytes from 199.91.189.38: icmp_seq=62 ttl=243 time=547.119 ms
64 bytes from 199.91.189.38: icmp_seq=63 ttl=243 time=467.761 ms
64 bytes from 199.91.189.38: icmp_seq=64 ttl=243 time=388.406 ms
64 bytes from 199.91.189.38: icmp_seq=65 ttl=243 time=557.122 ms
64 bytes from 199.91.189.38: icmp_seq=66 ttl=243 time=537.825 ms
64 bytes from 199.91.189.38: icmp_seq=67 ttl=243 time=102.962 ms
64 bytes from 199.91.189.38: icmp_seq=68 ttl=243 time=103.227 ms

http://www.as6453.net/

http://www.tatacommunications.com/

http://www.tatacommunications.com/ab...ontact_offices

http://en.wikipedia.org/wiki/Tata_Communications

http://www.ormuco.com/telecom-solutions/mpls/

Credit to Marishi-Ten




Now that you have a place to start investigating SE

[gigi_frana]

After a week of lag and 90k, yesterday was the first day I could play.
There is hope guys, hang in there.

[Jazellija]

Slingshot
New Zealand
Error 90000 disconnect after 2-3 minutes in an instance, either group or solo

Error 2002 when trying to get back into the character selection screen

Tethered with a mobile phone works no problem.

[Vithaar]

After a week of lag and 90k, yesterday was the first day I could play.
There is hope guys, hang in there.

same Gigi, maybe they fixed videotron, all day yesterday no problem

[Dragonovich]

- Tiscali/Talk Talk
- England,UK
- During peak times it seems. The connection cuts out once every 20-30mins.

[Istym]

UK - Sky broadband

4pm till around 11pm my icon flicks offline/online

GJ with the trace route ^^

[Konaim]

Charter Communications
Alabama, US
Always
----NPC appearance lag, major lag in cities with a lot of congestion. Even during off periods like 6a-10a still have 3-5 sec combat lag and quest NPC appearance. It will clear up for maybe 10 mins or so then back to the norm.

[Prometheus0110]

Time Warner Cable

New York City, NY

Always - I have 5,000+ Ping and rubber banding 100% of the time I've played. It translates into around 1.5 - 2 second skill lag, and massive rubber banding.

Side note - I pinged the IP address of the server where the game is run and it came back with 23ms.