375 replies to this topic

[BellatorMonk]

Yep still not paying for a 3rd party VPN/app w/e to solve a routing issue between TW-OVH-PGI. I dont pay $200 a month for Cable/Broadband to have to pay for VPN on top of that. It's silly.
Google just announced San Antonio just fell by the wayside for Fiber service in favor of Austin or Dallas..I forget..

[SoggyGorilla]

I am going to try to call them again, basically telia completely defends the fact that there is nothing wrong with there network and closed my case on me.

I just want to provide the last update to the ticket.



1/8/2015 10:52:31 AM E159740
Closing ticket.

This ticket was modified
-------------------------------------------------------------------------------
1/8/2015 10:52:18 AM E159740
Per Telia, the latency between DFW and Newark is within normal range for the distance. Notes indicate customer is seeing issues on ping times to game providers servers. There is no issue on the TWC network or peering connection. Closing ticket.

[Kin3ticX]

SoggyGorilla, on 31 January 2015 - 06:29 PM, said:

I am going to try to call them again, basically telia completely defends the fact that there is nothing wrong with there network and closed my case on me.

I just want to provide the last update to the ticket.



1/8/2015 10:52:31 AM E159740
Closing ticket.

This ticket was modified
-------------------------------------------------------------------------------
1/8/2015 10:52:18 AM E159740
Per Telia, the latency between DFW and Newark is within normal range for the distance. Notes indicate customer is seeing issues on ping times to game providers servers. There is no issue on the TWC network or peering connection. Closing ticket.

If it is not Telia, then maybe its a peer we cant see on the return route. We would need traceroutes from PGI to us. I think one of them was Tata Communications from one of the screenshots PGI Jon showed me.

TLDR: Some of us go to pgi through Telia and back on Tata Communications(Tcore). Some of us go to PGI through Tcore and back on Telia.

Its the only deduction I can come to from looking at Fierostats traceroute.

Edited by Kin3ticX, 31 January 2015 - 06:53 PM.

[LordSkippy]

BellatorMonk, on 31 January 2015 - 06:19 PM, said:

Google just announced San Antonio just fell by the wayside for Fiber service in favor of Austin or Dallas..I forget..

Must have been Dallas, because Austin was announced almost two years ago and sign-up for service in south Austin started last month. I'm in the "sign-ups coming soon" area, but on the other side of Austin from where they are signing up. One day, maybe, I can tell TWC to go suck it.

[SoggyGorilla]

Kin3ticX,

without a doubt is telia and the network problem they introduce. i can prove it with the following info.

ping and trace when connected to time warner (same pc, same location)



Ping 2 minutes later, after connecting my verizon LTE phone with a usb cable.




BOTH OF THESE PING TRACES HIT THE SAME ROUTER AT OVH. NWK-1-6K.

Now, you might ask how do you know this soggygorilla. Ill tell you.

because late one night i pinged the server as i always do when i started lagging, and saw somthing i have never seen, the hidden router responded with its name.



now




Now here is a picture of incoming data from telia to that router, in their USA weather map. (notice level3(verizon) hitting same router)

[Kin3ticX]

SoggyGorilla, on 31 January 2015 - 07:29 PM, said:

Kin3ticX,

without a doubt is telia and the network problem they introduce. i can prove it with the following info.

ping and trace when connected to time warner (same pc, same location)



Ping 2 minutes later, after connecting my verizon LTE phone with a usb cable.




BOTH OF THESE PING TRACES HIT THE SAME ROUTER AT OVH. NWK-1-6K.

Now, you might ask how do you know this soggygorilla. Ill tell you.

because late one night i pinged the server as i always do when i started lagging, and saw somthing i have never seen, the hidden router responded with its name.



now




Now here is a picture of incoming data from telia to that router, in their USA weather map. (notice level3(verizon) hitting same router)

My point is that since routings are asymmetric it is very likely the culprit is where you cant see it. From OVH to us. That and the problem hasnt been fixed and blaming Telia has done **** so far. Looking at the weathermap at 7:50pm pst, wherever I see Tata, I see 80% rather than a mere 15%. I dont have a router name to stamp it on but I know we are somehow going through tcore/tata.

examples:

http://i.imgur.com/RdTRdkE.png


Traceroute from OVH to me (about 1 month old)

http://i.imgur.com/MihBHLv.png
(im betting the one with the "????" would be locatable on the weathermap)


A recent traceroute from me to OVH.
http://i.imgur.com/WSHk5OO.png


On the weathermap Telia looks pretty good but tata has lots of red right now.

Edited by Kin3ticX, 31 January 2015 - 08:09 PM.

[SoggyGorilla]

yes so your coming from California.

looks like first step from telia to ovh happens in palo.

[Kin3ticX]

SoggyGorilla, on 31 January 2015 - 08:55 PM, said:

yes so your coming from California.

looks like first step from telia to ovh happens in palo.

So, your saying the reason only TWC is messed up is because of the way TWC enters the OVH network? I seem to forget the reason OVH is getting the pass here too.......

Edited by Kin3ticX, 31 January 2015 - 09:11 PM.

[SoggyGorilla]

Yes. my best guess is its some kind of problem with the translation effects of Telia traffic handing off to OVH. because time warner pays telia to route its traffic i have an avenue for support. however ovh is not a customer or has any service agreement to me or telia so nothing i can do on that end.

[NocturnalBeast]

SoggyGorilla, on 31 January 2015 - 11:16 PM, said:

Yes. my best guess is its some kind of problem with the translation effects of Telia traffic handing off to OVH. because time warner pays telia to route its traffic i have an avenue for support. however ovh is not a customer or has any service agreement to me or telia so nothing i can do on that end.

Yes, apparently Time Warner outsources their traffic to peer networks and some of the peer networks feel that they are not getting paid enough and are throttling Time Warner traffic as a protest.

[Clit Beastwood]

It might not even be that, if you do traceroutes via udp and icmp you get different results for each - I'm still kinda suspecting that MWO traffic is being treated as low priority - maybe they're treating the large port range + udp as bulk/p2p traffic? I've had days where the ping was fine and there were no issues, good hitreg/able to put down big numbers - usually playing at times when most people are asleep. Hard to tell the difference between that and an unladen switch tho.

[NocturnalBeast]

Fierostetz, on 01 February 2015 - 12:23 AM, said:

It might not even be that, if you do traceroutes via udp and icmp you get different results for each - I'm still kinda suspecting that MWO traffic is being treated as low priority - maybe they're treating the large port range + udp as bulk/p2p traffic? I've had days where the ping was fine and there were no issues, good hitreg/able to put down big numbers - usually playing at times when most people are asleep. Hard to tell the difference between that and an unladen switch tho.

Well, I had to turn off Pingzapper tonight, because I was constantly getting server disconnects and "unknown errors".

[BellatorMonk]

Another fun evening of 400-600ms matches:

9 dls-bb1-link.telia.net (213.248.93.189) 25.165 ms 23.690 ms 24.784 ms
10 nyk-bb1-link.telia.net (213.155.133.176) 74.745 ms 64.956 ms 65.773 ms
11 nyk-b2-link.telia.net (213.155.130.28) 64.752 ms 69.162 ms 69.860 ms
12 * * *
13 192.99.146.84 (192.99.146.84) 392.612 ms 416.188 ms 392.598 ms
14 198.27.73.231 (198.27.73.231) 119.502 ms 120.509 ms 119.640 ms
15 mwomercs.com (192.99.109.192) 120.737 ms !C 121.096 ms !C 117.563 ms !C

IP Location Canada Beauharnois Ovh Hosting Inc. ASN AS16276 OVH OVH SAS (registered Feb 15, 2001) Whois Server whois.arin.net IP Address 192.99.146.84

Time to start complaining:
http://www.ovh.com/us/index.xml

Edited by BellatorMonk, 01 February 2015 - 08:22 PM.

[Vanos]

Austin already has Google Fiber up and running in some areas. They are the major hub we would jump off of in San Antonio, sad face though if Dallas got it cause they didn't even submit a packet according to googles fiber.google.com/newcities page. They arn't even on the up coming possibilities list for Texas and we are still listed as possible in San Antonio.

[Vanos]

I encourge every one to post on this Time Warner Forums Thread to help TWC see impact, maybe together they will see the community needs help here. Thanks.

http://forums.timewa...-p/66270#M14677

Edited by Vanos, 02 February 2015 - 07:39 AM.

[Kin3ticX]

done

[NocturnalBeast]

MWO was largely unplayable yesterday, even with VPN. When I was able to get into a game, my ping was sub-100ms and very stable, but most of the time I was getting constant server disconnects. I know it was not my local connection, because I had no problems with Teamspeak. All I can think of is that maybe Super bowl web traffic was being heavily prioritized.

[Clit Beastwood]

yeah, with vpn I was getting a teensy bit of rubberbanding, hitreg was worse than it has been in a long time. AC20 seemed to work ok tho.

[Jon Cunningham]

Fierostetz, on 31 January 2015 - 09:59 AM, said:

As of this morning...

Just

Kin3ticX, on 31 January 2015 - 10:46 AM, said:

I noticed you don't go through telia when tracing to the website. Those vac3 hops are new for me too. Do you go through telia when you traceroute one of their gameservers? You can get the IP with wireshark or probably a console command.

The vac hops are part of the anti-DDoS solution. Occasionally traffic gets shunted there, evaluated, then passed back on to the system if it's valid.

[Clit Beastwood]

Jon Cunningham, on 02 February 2015 - 12:40 PM, said:

Just



The vac hops are part of the anti-DDoS solution. Occasionally traffic gets shunted there, evaluated, then passed back on to the system if it's valid.

Yeah that's what I'd read - do you guys have any idea if game traffic is being treated as bulk/low priority at your end? I'm just wondering if that might be why we get so much packet loss/latency on the last couple hops in Canada. Thanks for continuing to reply to the thread