184 replies to this topic

[RichAC]

ShinVector, on 05 January 2014 - 06:33 AM, said:

You can use notepad or what ever to save the info from WinMTR record the date and time. This is to create a log.
When your high ping issue hits, its should reveal at which hop the deviation is and where the problem starts.

Taking note.
Usually the first hop is your local network. If its goes bad there.. There is likely a local LAN issue.
Next few hops will be the local loop to your ISP..
After your ISP it will moving to other ISPs through the internet to reach MWO servers.

For eg.
I had own connectivity problem several months back which I was too lazy to troubleshoot and dismissed bad connectivity to Canada.
When I start troubleshooting I realise I was having latency spikes and packets loss even my own ISP's gateway IP on the local loop portion.
Problem found to be due to a loose optical fibre cable at the fibre terminate point in my living room where my FTHT optical fibre modem connects too. (Maybe one of cats did it who knows.)

Essentially want to avoid troubleshooting aimlessly if the problem is in your own house.

3000ms is kinda weird and extreme.. Try to log it down and know where the spikes are coming from as log from winMTR.

Ok shin, now we are all just lazy and the trouble is in our own house or local network....lol The more research i do on this issue, and the more persistent I am, the more it seems like its out of our hands and something else is going on.

a traceroute is good enough to tell these things. Its what these admins ask for. Spiking is spiking...its not a steady 800 ping for some of these europeans if the traceroute is showing them 250 or thats what their constant ping used to be. constant spiking like that would also show in the traceroute with an asterik.

I saw the guy from the phiillipines John E. quoted in this thread. He had a steady 800 ping, he wasn't even spiking. It was like he was being throttled. and spiking from in the 100s...to the thousands is already a non issue, beause the guy already has ping in the high 100s to being with!

Care to comment on the links i posted, which show this exact issue also happened back in january immediately after MWO had some downtime and released a new patch?

Apparenlty MWO has that many players that play immediately after a patch that it overloads most ISP's circuits? I'm finding it harder and harder to believe man. Something is fishy about this.

I still say either something is going on to try and level everybodies ping in the game actually doing more harm. maybe its a jealous malcious user... anything is possible to me at this point. Or pgi doesn't pay enough for bandwidth or dont' rate as a small company.

Or its just an issue routing to canada for most isp's, especially during the winter months when there is more load.

I don't think there is much we can do about it on our end. I doubt the admin from germany is going to reply back to me...and this time that server isn't actually owned by verizon. SO me constanlty harassing verizon, like believe me I've been doing, isn't really going to get me anywhere. The circuit on that router in question now is an internap circuit and the router is owned by tiscali/tinet.net.

Either way its mind boggling and this is the only game anybody has issues with. maybe they should host in toronto, I dunno what would fix this.

Edited by RichAC, 05 January 2014 - 07:01 AM.

[John E]

my ping was 260ms to begin with rich, also i got a reply from the pgi staff and they are aware of this issue so I'm leaving it up to them. I doubt we will be able to do anything about it.

[RichAC]

John E, on 05 January 2014 - 08:06 AM, said:

my ping was 260ms to begin with rich, also i got a reply from the pgi staff and they are aware of this issue so I'm leaving it up to them. I doubt we will be able to do anything about it.

ya that other Phillipino guy you quoted I seen in game. He said he had same ping too around 250 usually. Now his ping was 800!

After reading the MWO thread from back in january. It seems like folks in the northeast with verizon finally got verizon to fix circuit k on verizons border/link router or w/e its called that routes to tinet.net:

tinet-gw.customer.alter.net [152.179.72.122]

alter.net is owned by verizon, which like with me, verizon first tried to deny then had to admit. I originally believed the issue was because this router was sending me to germany. But the route doesn't change when I have 20 ping.

In january, the knowledgeable mwo players that used to play and were customers of verizon were possibly able to pressure verizon to fixing that circuit on the router. Because that server was actualy owned by verizon themselves. It seems like none of them even play the game anymore so maybe this isn't the case. I wish they were still around to reply.
But now my problem is, my ping starts to go bad with the next link router further up the line in the same route...which is:

internap-gw.ip4.tinet.net [77.67.70.94]

tinet.net is owned by German company. Which is now the router that first starts to go bad, and who i emailed for help. I can't tell Verizon they own this server and demand that administer it themselves.

And I'm no network expert, but it seems what happened is the issue got passed up the line. SO if this router increases bandwidth for that circuit, if the problem is indeed load, and which according to verizon internap is the circuit people for that rouer, then internap will be the ones overloaded next.

And then its back to square one....it will be on PGI's isp to fix the issue. Which i'm starting to agree with you is maybe the root of the problem to begin with. I think Canada's routing just sucks in general. speedtests and raceroutes to other vancouver servers are no better.

But then again those are not game servers and there might be a difference. The only other game I play with a gameserver in canada, is quakelive and there server is in toronto which i get decent ping on. or in BF3.

Maybe someone else can comment to correct me.

Edited by RichAC, 05 January 2014 - 09:26 AM.

[RichAC]

actually this is interesting. I went to browse battlefield 3 servers. and set filter to canadian servers. It seems There aren't any! hahah.

I could of sworn there used to be. dozens for USA and other countries come up. This confirms for me there is a problem connecting to canada, and noone wants to host their game server there.

Edited by RichAC, 05 January 2014 - 09:36 AM.

[ShinVector]

RichAC, on 05 January 2014 - 06:45 AM, said:

Ok shin, now we are all just lazy and the trouble is in our own house or local network....lol The more research i do on this issue, and the more persistent I am, the more it seems like its out of our hands and something else is going on.

Woo... Hold on there.. How does myself sharing experience on solving my own connectivity problem in the past become me calling everyone lazy ? I assume you were joking about that but still people who aren't following the thread might missed understand that statement.

What I am saying every issue a per any other problem needs to be investigated separately.
One cannot simply cannot blame everything on a patch like some have been doing so, logically in the troubleshooting process it makes no sense nor does help to reach any conclusion.

Please do not assume someone in the Philippines still going to face the same problem you do. The symptoms are already different +50ms latency increase versus +800ms.
If the dude had put his traceroute or winmtr of his issue down here where the problem is occurring we could probably tell more.

Your problem based on information gather is most likely due to link congestion during peak hours between tinet-internap gateway. Only they have the information on the exact cause and can solve the problem.
However other like myself might consider a +50ms different trivial since we are still able to play this at 4-6 times the latency. Either chase the ISP to solve the issue or live with it or change to a better ISP.

Yes. Agreed on the Canada part.. Probably a bad location to put your only game server at. This has been mentioned in the past.. then they talk about us having to separate account in said regional servers.. lol.

[Kh0rn]

Alright I re did the test with a program that shows me the hops as you described the first few worked perfectly but the last one had some serious issues. Guess it confirms its nothing on my side every other game shows 260 ping which is perfect but MWO is still a no no. But as John said PGI is working on it I'm leaving it too them now.

[ShinVector]

Kh0rn, on 06 January 2014 - 12:25 AM, said:

Alright I re did the test with a program that shows me the hops as you described the first few worked perfectly but the last one had some serious issues. Guess it confirms its nothing on my side every other game shows 260 ping which is perfect but MWO is still a no no. But as John said PGI is working on it I'm leaving it too them now.

You can post it here if you need deciphering.
Do note we can't only icmp up to a certain point PGI is probably blocking pings to their game servers itself.

[RichAC]

ShinVector, on 05 January 2014 - 05:23 PM, said:

Woo... Hold on there.. How does myself sharing experience on solving my own connectivity problem in the past become me calling everyone lazy ? I assume you were joking about that but still people who aren't following the thread might missed understand that statement.

What I am saying every issue a per any other problem needs to be investigated separately.
One cannot simply cannot blame everything on a patch like some have been doing so, logically in the troubleshooting process it makes no sense nor does help to reach any conclusion.

Please do not assume someone in the Philippines still going to face the same problem you do. The symptoms are already different +50ms latency increase versus +800ms.
If the dude had put his traceroute or winmtr of his issue down here where the problem is occurring we could probably tell more.

Your problem based on information gather is most likely due to link congestion during peak hours between tinet-internap gateway. Only they have the information on the exact cause and can solve the problem.
However other like myself might consider a +50ms different trivial since we are still able to play this at 4-6 times the latency. Either chase the ISP to solve the issue or live with it or change to a better ISP.

Yes. Agreed on the Canada part.. Probably a bad location to put your only game server at. This has been mentioned in the past.. then they talk about us having to separate account in said regional servers.. lol.

Well if you followed the thread you will see I was calling these guys crazy for thinking it had to do with PGI or their patch. But then i found those threads in january filled with people who no longer play the game, feeling the same ways.

Now it could be a placebo and its just that our ISP's can't handle the load at this time of year, and that maybe it was colder in january last year....lmao, so more people coincidentally went online at that time. and its the ISP's that are throttling us, and not PGI for more level ping among players.....

I have been contacting verizon, I emailed the tinet admin of the router with the bad internap cirucit. I agree its poor business practice for verizon to ask me to do that, and I would have to be crazy to expect any sort of reply....

Time Warner users in my area have perfect ping to MWO....

But I'm no network expert, and apparently neither was the lvl 3 guy at verizon, who now has all the lvl 1 and 2 guys, including the last guy i talked to, giving me the runaround, and I can't really rule anything out anymore...

Its way easier to hit lights with my ppcs at 20 ping then it is when my ping is spiking from 100-150 lol. But Its not so much the 100 ping, since thats avg for this game. Its just the random crashes and disconnects that happen when my ping gets that high. And i really don't care how rare they occur, its the fact I know theres a chance of it happening because of my connection that bothers me.

Edited by RichAC, 06 January 2014 - 04:18 AM.

[RichAC]

Also if you read one of the links I posted regarding the european players problems in january, not verizon customers.

You will see it was PGI who replied back, addressed the issue and fixed it... something about the crynetwork layer.

If its the case, that Us Americans are crashing or losing connection though, because PGi wants us all at the same ping as europeans, its only driving people away from this already dying game.

maybe its just the fact vancouver is a horrible place to host this game, and us Americans are spoiled.

Edited by RichAC, 06 January 2014 - 05:04 AM.

[DragonsFire]

Just to add to this thread, I'm looking to see about getting some information on some latency increases that I've noticed as of late during the hours of 8 PM - 12 AM EST. The increase I am specifically tracking comes in here:

Hop PL% Min Max Avg Host Name / [IP]
12 0.2 37 102 52 if-5-0-0-5.core4.TNK-Toronto.as6453.net [63.243.172.25]

13 6.9 92 136 98 ix-4-0-3.core4.TNK-Toronto.as6453.net [216.6.49.2]

14 7.0 92 447 102 border1.te9-1-bbnet2.tor001.pnap.net [70.42.24.196]

Hop 13 (216.6.49.2), which is owned by Tata Communications, shows a consistent increase in latency over hop 12, which is then carried to the datacenter border and from there to the MWO servers.

What I'm trying to identify is if anyone else is seeing this latency increase as well and traversing this same path. There is also a 7% packet loss rate happening there, which is further complicating matters. If any of you out there are having an issue with latency and could perform a trace to verify your path and post the results here if they are similar, that would be greatly appreciated.

PGI can't do anything directly as this is outside of the datacenter but with enough information either they or the datacenter provider could lean on Tata communications to make some corrections.

[RichAC]

DragonsFire, on 06 January 2014 - 06:57 AM, said:

Just to add to this thread, I'm looking to see about getting some information on some latency increases that I've noticed as of late during the hours of 8 PM - 12 AM EST. The increase I am specifically tracking comes in here:

Hop PL% Min Max Avg Host Name / [IP]
12 0.2 37 102 52 if-5-0-0-5.core4.TNK-Toronto.as6453.net [63.243.172.25]

13 6.9 92 136 98 ix-4-0-3.core4.TNK-Toronto.as6453.net [216.6.49.2]

14 7.0 92 447 102 border1.te9-1-bbnet2.tor001.pnap.net [70.42.24.196]

Hop 13 (216.6.49.2), which is owned by Tata Communications, shows a consistent increase in latency over hop 12, which is then carried to the datacenter border and from there to the MWO servers.

What I'm trying to identify is if anyone else is seeing this latency increase as well and traversing this same path. There is also a 7% packet loss rate happening there, which is further complicating matters. If any of you out there are having an issue with latency and could perform a trace to verify your path and post the results here if they are similar, that would be greatly appreciated.

PGI can't do anything directly as this is outside of the datacenter but with enough information either they or the datacenter provider could lean on Tata communications to make some corrections.

and then i bet the load gets pushed on internap and then PGI or whoever owns their server will have to pay them more money I bet...lol Forget canada altogther imo, host the server in washington dc or nyc.

Edited by RichAC, 06 January 2014 - 02:57 PM.

[ShinVector]

DragonsFire, on 06 January 2014 - 06:57 AM, said:

Just to add to this thread, I'm looking to see about getting some information on some latency increases that I've noticed as of late during the hours of 8 PM - 12 AM EST. The increase I am specifically tracking comes in here:

Hop PL% Min Max Avg Host Name / [IP]
12 0.2 37 102 52 if-5-0-0-5.core4.TNK-Toronto.as6453.net [63.243.172.25]

13 6.9 92 136 98 ix-4-0-3.core4.TNK-Toronto.as6453.net [216.6.49.2]

14 7.0 92 447 102 border1.te9-1-bbnet2.tor001.pnap.net [70.42.24.196]

Hmmm.. I am at GMT+8 so, your 8pm should be my 9am.
7% would be certainly bad for gaming...
Did you get it consistently after a few tries ? You should for at least 100-200 packets.

[RichAC]

you can try to email them. all whois shows is
trouble@as6453.net maybe you can ask your isp if they have another contact for them. I have yet to get emailed back on the troubled circuit on my route. I mean is it even common for them too? it sounds ridiculous that is our only option...

Edited by RichAC, 06 January 2014 - 09:23 PM.

[DragonsFire]

ShinVector, on 06 January 2014 - 07:56 PM, said:

Hmmm.. I am at GMT+8 so, your 8pm should be my 9am.
7% would be certainly bad for gaming...
Did you get it consistently after a few tries ? You should for at least 100-200 packets.

I had a 10000 query session set up in PingPlotter which gave me the 7% loss over the full course of the query unfortunately. The loss definitely does make gaming during the increased latency difficult. I can deal with (and HSR handles it as well) the latency increase but the loss just throws it all out the window so I just generally don't play until it's no longer present as it's too frustrating otherwise.

[ShinVector]

DragonsFire, on 06 January 2014 - 09:22 PM, said:

I had a 10000 query session set up in PingPlotter which gave me the 7% loss over the full course of the query unfortunately. The loss definitely does make gaming during the increased latency difficult. I can deal with (and HSR handles it as well) the latency increase but the loss just throws it all out the window so I just generally don't play until it's no longer present as it's too frustrating otherwise.

RichAc, said it... Other people have said.. Going to repeat it again... If only PGI had game servers in other more well connected countries..

[RichAC]

ShinVector, on 06 January 2014 - 09:35 PM, said:

RichAc, said it... Other people have said.. Going to repeat it again... If only PGI had game servers in other more well connected countries..

I ping decent to toronto, canada. There is one guy in game that has a fast connection in toronto and gets 6 ping to the game server haha. But most of them get ping in the 20s like I used to. The only ISPs that ping good seem to be in northeast, like toronto, ohio, vermont, time warner in nyc. Northeast connects best to vancouver it seems.

Brasillians in BF3 all host their servers in nyc, it seems that phillipino guy had to go through ny to get to vancouver as well... I think toronto is the best canadian choice.

what part of North America do you ping best to Shin?

Edited by RichAC, 07 January 2014 - 04:33 AM.

[RichAC]

Shin, have you actually identified where this server is?

Cause utrace shows it as being in toronto, people in game have said montreal, I thought it was in vancouver all this time. I don't remember why I thought that. I think i was totally wrong about that. The server is probably in toronto!!!!

this is the server i have to connect to, to log in and play a game with client:

relay-1.mwtactics.com [70.42.29.65]

http://www.ip-adress...s/mwtactics.com

http://en.utrace.de/?query=70.42.29.65

If this server is in toronto then I dont' know what to say. I speedtest great to toronto and play on quakelive toronto servers with great ping... I just did some speedtests at speedtest.net to montreal and toronto servers. I get anywhere from 20-50 ping depending. not 100 like i'm getting in game...

;Once again I'm now thinking it has more to do with their allocated bandwidth rather then my ISP.

Or maybe my ip address goes on the list to get throttled to european pings for level play? I'm thinking again maybe it is true PGI tries to keep everyone at 100 ping. Maybe thats why sometimes when I change to a random ip i'm immediately back to the 25 ping until throttled again after 24 hours because its an ip or verizon fios range the routers haven't seen yet?..... but then that doesn't explain why some users, like the toronot users get constant 20 ping. Maybe its only certain ISP's that help to allocate PGI's bandwidth or thortle their pings?

this is even more mind boggling then before...All this time I'm saying the server is in vancouver, which made me think its my ISP, noone ever corrected me and it turns out the server is actually in toronto! which i ping great to on any other server?

I'm at a loss for words....

Edited by RichAC, 07 January 2014 - 07:51 AM.

[VerTicaL]

Damn I hope they fixed it...seems like the best ones

fingers crossed...

[DragonsFire]

RichAC, on 07 January 2014 - 06:52 AM, said:

snip

Server is definitely in Toronto. The datacenter there is owned by Internap (or perhaps another company but the IP space belongs to Internap) and incoming connections are peered based on various routing decisions happening at the edge. Judging by your previous post, your latency is incurred on an Internap router a couple of hops before the server itself.

It could be that their is a load balancing configuration that's been applied to one of these edge routers that is not working properly and thus excessively loading one TenGig link while the other remains underutilized. Unfortunately without direct access to these routers that's the best guess I can make.

What's important to note though is that this is not something directly attributed to PGI as this latency is being incurred a few hops prior to their servers. Further, while they can certainly make requests for it to be looked into, datacenter work generally won't get much traction unless there is enough hard evidence to warrant potentially service impacting work due to SLA (Service Level Agreements) that are in place with other customers at the site.

The best thing we can do is to try and document when you see the increases, if it's constant, what your traces to the servers look like, etc. That will help to be able to build sufficient evidence that there is something that either the datacenter group needs to look at, the Service Provider needs to look at, or both.

[RichAC]

The bad hop in question is actually a tinet router that connects me to internap. The name at the beginnig of the domain name is actually the name of the network or circuit that link/border/edge router is connecting me to. The name at the end of the domain name is whose router it is. The lvl 3 guy at verizon was also confused about this...which is why he thought the first non local link/border/edge router was owned by tinet and not verizon who is actually alter.net. (internap is pnap.net) Its actually really straight forward and common sensical and I don't know why it goes right over the head of so called professionals in the industry.


But anyway, I emailed the admin of that problem router. I politely explained the situation and sent him bunch of good and bad traceroutes. But I think I'd have to be an {Dezgra} to expect any kind of reply back.

Whats baffling to me now is that i get the same ping in speedtests to toronto at speedtest.net, that i get on a toronto quakelive server, which is more then half of what I get connecting to PGI. I'm not sure why this is. and I can't really rule anybody out, because I don't know what kind of special arrangements these companies have with each other.

I'll harass verizon some more monday, someone else posted about mwo on their forums lately also. I'll call monday and say i have gotten no reply back from the admin of the router in question, and that timewarner cable and comcast satellite are getting 20 ping and I'm getting 100 ping on fios...is there anything else that can be done. Or is there something else you suggest I do or say?

A verizon guy called me and said he played MWO, he even sounded like someone I've played with on teamspeak before lol, he said he lives in florida, and he was saying that either PGI isn't paying for enough bandwidth, or that circuit on that router needs to be upgraded for more load.

Call me a conspiracy theorist, but I wonder if most of these companies know the game, and they feel 100 ping is giving the game a level playing field, as an excuse for themselves to offload bandwidth, and that they feel I shoudln't be complaining about it..

Edited by RichAC, 07 January 2014 - 05:01 PM.