332 replies to this topic

[Oderint dum Metuant]

AUSwarrior24, on 04 September 2014 - 02:53 AM, said:

Okay, may as well.

Tracing route to astralinternet.com [198.72.103.155]
over a maximum of 30 hops:
  1	 1 ms	 2 ms	 2 ms  192.168.0.1
  2	48 ms	45 ms	42 ms  250.2.2.123.network.m2core.net.au [123.2.2.250]
  3	44 ms	45 ms	71 ms  be2-v417.bsr02.melbpri.vic.m2core.net.au [123.2.
2.65]
  4	56 ms	47 ms	59 ms  bundle-ether11.win24.melbourne.telstra.net [165.
228.21.137]
  5	68 ms	69 ms	54 ms  bundle-ether6.win-core1.melbourne.telstra.net [2
03.50.11.58]
  6	55 ms	57 ms	58 ms  bundle-ether12.ken-core10.sydney.telstra.net [20
3.50.11.122]
  7	59 ms	78 ms	67 ms  Bundle-ether18.pad-gw2.sydney.telstra.net [203.5
0.13.82]
  8	82 ms	70 ms	67 ms  bundle-ether1.sydp-core01.sydney.reach.com [203.
50.13.46]
  9   203 ms   191 ms   194 ms  i-0-1-1-3.istt-core02.bx.telstraglobal.net [202.
84.140.82]
10   220 ms   191 ms   207 ms  i-0-4-0-0.paix02.bi.telstraglobal.net [202.84.25
1.58]
11   222 ms   222 ms   216 ms  te0-0-0-3.ccr21.sjc04.atlas.cogentco.com [154.54
.11.157]
12   220 ms   228 ms   243 ms  be2015.ccr21.sfo01.atlas.cogentco.com [154.54.7.
173]
13   245 ms   257 ms   245 ms  be2132.ccr21.mci01.atlas.cogentco.com [154.54.30
.54]
14	 *	  354 ms   302 ms  be2156.ccr41.ord01.atlas.cogentco.com [154.54.6.
86]
15   291 ms   295 ms   268 ms  be2079.ccr21.yyz02.atlas.cogentco.com [154.54.27
.182]
16   301 ms   281 ms   284 ms  be2090.ccr21.ymq02.atlas.cogentco.com [154.54.30
.206]
17   393 ms   364 ms   351 ms  te8-4.cl-core06.cogent.mtl.iweb.com [38.104.154.
202]
18   405 ms   375 ms   434 ms  ae1.cr7.mtl.iweb.com [184.107.1.210]
19   392 ms   427 ms   392 ms  po22.cr3.mtl.iweb.com [184.107.1.2]
20   357 ms   373 ms   351 ms  te9-3.dr4.mtl.iweb.com [67.205.127.85]
21   364 ms   374 ms   350 ms  www.astralinternet.com [198.72.103.155]
Trace complete.

 
Tracing route to hostutopia.com [108.163.188.198]
over a maximum of 30 hops:
  1	 1 ms	 1 ms	 1 ms  192.168.0.1
  2	41 ms	43 ms	44 ms  250.2.2.123.network.m2core.net.au [123.2.2.250]
  3	71 ms	56 ms	61 ms  be2-v417.bsr02.melbpri.vic.m2core.net.au [123.2.
2.65]
  4	52 ms	47 ms	71 ms  bundle-ether11.win24.melbourne.telstra.net [165.
228.21.137]
  5	56 ms	54 ms	45 ms  bundle-ether6.win-core1.melbourne.telstra.net [2
03.50.11.58]
  6	64 ms	87 ms	70 ms  bundle-ether12.ken-core10.sydney.telstra.net [20
3.50.11.122]
  7	69 ms	60 ms	64 ms  Bundle-ether18.pad-gw2.sydney.telstra.net [203.5
0.13.82]
  8	63 ms	55 ms	56 ms  bundle-ether1.sydp-core01.sydney.reach.com [203.
50.13.46]
  9   201 ms   215 ms   203 ms  i-0-3-0-6.paix-core01.bx.telstraglobal.net [202.
84.140.194]
10   199 ms   199 ms   202 ms  i-0-5-0-2.paix02.bi.telstraglobal.net [202.40.14
9.114]
11   214 ms   209 ms   235 ms  te0-0-0-3.ccr21.sjc04.atlas.cogentco.com [154.54
.11.157]
12   254 ms   229 ms   226 ms  be2015.ccr21.sfo01.atlas.cogentco.com [154.54.7.
173]
13   264 ms   246 ms   284 ms  be2132.ccr21.mci01.atlas.cogentco.com [154.54.30
.54]
14   288 ms   259 ms   281 ms  be2156.ccr41.ord01.atlas.cogentco.com [154.54.6.
86]
15   284 ms   274 ms   290 ms  be2079.ccr21.yyz02.atlas.cogentco.com [154.54.27
.182]
16   283 ms   280 ms   332 ms  be2090.ccr21.ymq02.atlas.cogentco.com [154.54.30
.206]
17   355 ms   371 ms   360 ms  38.122.42.122
18   402 ms   398 ms   399 ms  po21.dr13.mtl.iweb.com [184.107.1.246]
19   374 ms   398 ms   366 ms  hostutopia.com [108.163.188.198]
Trace complete.

Only looked through the first one quickly at work but it looks normal; which is no surprise as it's a different address.

[Oderint dum Metuant]

Nayonac, on 04 September 2014 - 03:13 AM, said:

So we have 2 problems here
1) De-syncs for everyone
2) Higher ping's for some people / lower ping's for other people...

So looking at problem 2)
They have moved Datacentre, the servers are not in the same physical location now and thus your traffic is now going to different locations, which means your internet traffic is now going a different route.

So for those that don't understand why you have a higher or a lower ping lets say you have a job.
That takes you 30 minute to get to, suddenly the Job that takes 30 minutes to get to relocates to the location 60 minutes away, so now you have to travel 60 minutes instead of 30, sure you might be able to change your route to work and optimize your journey but it will not be the same one you had before. So what we are saying is that your traffic is not going via the same route as it used to and thus change has occurred.

You can perform a trace route and see this change(if you knew the route before), however your traffic is governed internally by your ISP on its own network, once this traffic leaves the network onto another network, that traffic is governed by the rules of the next ISP (normally done via a weighting system of preferences on peers) so your ISP will have no say over where it goes from there. The same goes for PGI, whoever they buy bandwidth from will have a choice of which way the traffic leaves there network (which might mean your traffic back comes along a different route to the one it went on) once it has left they have no power over what happens to it once it has left there network.
So your never going to get a nice low ping unless your ISP has a direct peer to the network PGI are hosted on.
So your choices are change ISP and hope they go via a better route or ask your current one if they and shift there route preferences so traffic takes a different route (probably never happen unless they see a cost benefit), or move closer to the Datacentre... PGI depending if they just buy bandwidth or run there own routers could buy in alternate sources or get free peers to other ISP's which may preference your route better (again affected by the out bound route of your ISP and luck of the draw on which way your traffic gets routed

Only one point of contention here your ISP in this case Telstra can indeed influence how the data travels outside their network if there is indeed a problem with the routing.

It happens quite often here in Europe albeit with bigger games (LOL and WoW) for example have had routing issues before thanks to dodgy routing from Telia but the problem needs to be highlighted to your ISP and in turn they will highlight it to their tier partner.

That's not to say it doesn't take along time to get resolved.

[N0MAD]

AUSwarrior24, on 04 September 2014 - 02:53 AM, said:

Okay, may as well.

Tracing route to astralinternet.com [198.72.103.155]
over a maximum of 30 hops:
  1	 1 ms	 2 ms	 2 ms  192.168.0.1
  2	48 ms	45 ms	42 ms  250.2.2.123.network.m2core.net.au [123.2.2.250]
  3	44 ms	45 ms	71 ms  be2-v417.bsr02.melbpri.vic.m2core.net.au [123.2.
2.65]
  4	56 ms	47 ms	59 ms  bundle-ether11.win24.melbourne.telstra.net [165.
228.21.137]
  5	68 ms	69 ms	54 ms  bundle-ether6.win-core1.melbourne.telstra.net [2
03.50.11.58]
  6	55 ms	57 ms	58 ms  bundle-ether12.ken-core10.sydney.telstra.net [20
3.50.11.122]
  7	59 ms	78 ms	67 ms  Bundle-ether18.pad-gw2.sydney.telstra.net [203.5
0.13.82]
  8	82 ms	70 ms	67 ms  bundle-ether1.sydp-core01.sydney.reach.com [203.
50.13.46]
  9   203 ms   191 ms   194 ms  i-0-1-1-3.istt-core02.bx.telstraglobal.net [202.
84.140.82]
10   220 ms   191 ms   207 ms  i-0-4-0-0.paix02.bi.telstraglobal.net [202.84.25
1.58]
11   222 ms   222 ms   216 ms  te0-0-0-3.ccr21.sjc04.atlas.cogentco.com [154.54
.11.157]
12   220 ms   228 ms   243 ms  be2015.ccr21.sfo01.atlas.cogentco.com [154.54.7.
173]
13   245 ms   257 ms   245 ms  be2132.ccr21.mci01.atlas.cogentco.com [154.54.30
.54]
14	 *	  354 ms   302 ms  be2156.ccr41.ord01.atlas.cogentco.com [154.54.6.
86]
15   291 ms   295 ms   268 ms  be2079.ccr21.yyz02.atlas.cogentco.com [154.54.27
.182]
16   301 ms   281 ms   284 ms  be2090.ccr21.ymq02.atlas.cogentco.com [154.54.30
.206]
17   393 ms   364 ms   351 ms  te8-4.cl-core06.cogent.mtl.iweb.com [38.104.154.
202]
18   405 ms   375 ms   434 ms  ae1.cr7.mtl.iweb.com [184.107.1.210]
19   392 ms   427 ms   392 ms  po22.cr3.mtl.iweb.com [184.107.1.2]
20   357 ms   373 ms   351 ms  te9-3.dr4.mtl.iweb.com [67.205.127.85]
21   364 ms   374 ms   350 ms  www.astralinternet.com [198.72.103.155]
Trace complete.

 
Tracing route to hostutopia.com [108.163.188.198]
over a maximum of 30 hops:
  1	 1 ms	 1 ms	 1 ms  192.168.0.1
  2	41 ms	43 ms	44 ms  250.2.2.123.network.m2core.net.au [123.2.2.250]
  3	71 ms	56 ms	61 ms  be2-v417.bsr02.melbpri.vic.m2core.net.au [123.2.
2.65]
  4	52 ms	47 ms	71 ms  bundle-ether11.win24.melbourne.telstra.net [165.
228.21.137]
  5	56 ms	54 ms	45 ms  bundle-ether6.win-core1.melbourne.telstra.net [2
03.50.11.58]
  6	64 ms	87 ms	70 ms  bundle-ether12.ken-core10.sydney.telstra.net [20
3.50.11.122]
  7	69 ms	60 ms	64 ms  Bundle-ether18.pad-gw2.sydney.telstra.net [203.5
0.13.82]
  8	63 ms	55 ms	56 ms  bundle-ether1.sydp-core01.sydney.reach.com [203.
50.13.46]
  9   201 ms   215 ms   203 ms  i-0-3-0-6.paix-core01.bx.telstraglobal.net [202.
84.140.194]
10   199 ms   199 ms   202 ms  i-0-5-0-2.paix02.bi.telstraglobal.net [202.40.14
9.114]
11   214 ms   209 ms   235 ms  te0-0-0-3.ccr21.sjc04.atlas.cogentco.com [154.54
.11.157]
12   254 ms   229 ms   226 ms  be2015.ccr21.sfo01.atlas.cogentco.com [154.54.7.
173]
13   264 ms   246 ms   284 ms  be2132.ccr21.mci01.atlas.cogentco.com [154.54.30
.54]
14   288 ms   259 ms   281 ms  be2156.ccr41.ord01.atlas.cogentco.com [154.54.6.
86]
15   284 ms   274 ms   290 ms  be2079.ccr21.yyz02.atlas.cogentco.com [154.54.27
.182]
16   283 ms   280 ms   332 ms  be2090.ccr21.ymq02.atlas.cogentco.com [154.54.30
.206]
17   355 ms   371 ms   360 ms  38.122.42.122
18   402 ms   398 ms   399 ms  po21.dr13.mtl.iweb.com [184.107.1.246]
19   374 ms   398 ms   366 ms  hostutopia.com [108.163.188.198]
Trace complete.

So mate were you been routed thru EU to MWO?.
The traces you have supplied to addys in Montreal as you see are routed thru the US to canada.

[Egomane]

N0MAD, on 04 September 2014 - 06:49 AM, said:

So mate were you been routed thru EU to MWO?.
The traces you have supplied to addys in Montreal as you see are routed thru the US to canada.

How many times do we have to explain it?
Routing is not based on physical distance between targets. Physical existence in the real world is mostly irrelevant!

You can have two adresses behind the same router. Physically right next to each other!
If you send data to number one, the package might use the shortest physical route possible.
If you send data to number two, it will take a detour two times around the world.

[Oderint dum Metuant]

Egomane, on 04 September 2014 - 07:29 AM, said:

How many times do we have to explain it?
Routing is not based on physical distance between targets. Physical existence in the real world is mostly irrelevant!

You can have two adresses behind the same router. Physically right next to each other!
If you send data to number one, the package might use the shortest physical route possible.
If you send data to number two, it will take a detour two times around the world.

#lostcause

Classic case of I think I know vs I know I know

Edited by DV McKenna, 04 September 2014 - 07:34 AM.

[VanillaG]

N0MAD, on 04 September 2014 - 06:49 AM, said:

So mate were you been routed thru EU to MWO?.
The traces you have supplied to addys in Montreal as you see are routed thru the US to canada.

Those sites are hosted in a different data center and use a different transit provider, Cogent, than the ones that MWO is using. Looking at all of my traces MWO is being hosted by OVH so you would get different routes because they are on totally different providers.

My guess is that issue that Aussie players are seeing is that some segments of the their ISP network are getting routed to Europe because OVH is a large European based host. Their routers most likely see an OVH IP address and assume Europe so try to route through there.

[Beren Valari]

Mahws, on 02 September 2014 - 08:54 PM, said:

Any word on whether other countries are effected? In Japan my ping has increased by 80 since the server relocation.

Edit:
Just did a trace, going through italy and france to reach mwomercs.com. :/

N0MAD, on 03 September 2014 - 01:53 AM, said:

My trace from Brisbane is near identical to yours once it hits Sydney with almost identical pings.
Telstra does not set your rout mate, we all goto the Telstra Sydney hub from there the hubs do their work, they will only deviate if they are forced to but its not Telstra setting the routes.
Telstra may eat babies but they dont dictate your internet routes once they hit their international hub, you do that (if you want to)

Microsoft Windows [Version 6.1.7601]
Copyright (c) 2009 Microsoft Corporation. All rights reserved.

C:\Users\Luke>tracert mwomercs.com

Tracing route to mwomercs.com [192.99.109.192]
over a maximum of 30 hops:

1 3 ms 4 ms 4 ms DD-WRT [192.168.11.1]
2 * * * Request timed out.
3 15 ms 62 ms 27 ms 218.53.148.93
4 24 ms 38 ms 46 ms 58.229.28.9
5 33 ms 36 ms 50 ms 58.229.119.205
6 22 ms 18 ms 24 ms 1.255.26.246
7 49 ms 34 ms 19 ms 58.229.14.70
8 158 ms 160 ms 136 ms 58.229.15.202
9 * * * Request timed out.
10 335 ms 364 ms 351 ms rbx-g1-a9.fr.eu [178.33.100.238]
11 353 ms 347 ms 360 ms 37.187.36.131
12 * * * Request timed out.
13 369 ms 363 ms 398 ms mwomercs.com [192.99.109.192]

Trace complete.

C:\Users\Luke>

I'm in South Korea and getting routed through the same ****** French server. So now tell me again how its an AU ISP thats causing the issue. I had a 220ms ping before the move and now average 360ms. That's just unacceptable. Especially if they've moved to a cheaper server after I've payed them a bunch of cash. At the very least PGI should set up a VPN or VPS for us in AU/NZ and East Asia.

[Oderint dum Metuant]

Beren Valari, on 04 September 2014 - 07:57 AM, said:

Microsoft Windows [Version 6.1.7601]
Copyright (c) 2009 Microsoft Corporation. All rights reserved.

C:\Users\Luke>tracert mwomercs.com

Tracing route to mwomercs.com [192.99.109.192]
over a maximum of 30 hops:

1 3 ms 4 ms 4 ms DD-WRT [192.168.11.1]
2 * * * Request timed out.
3 15 ms 62 ms 27 ms 218.53.148.93
4 24 ms 38 ms 46 ms 58.229.28.9
5 33 ms 36 ms 50 ms 58.229.119.205
6 22 ms 18 ms 24 ms 1.255.26.246
7 49 ms 34 ms 19 ms 58.229.14.70
8 158 ms 160 ms 136 ms 58.229.15.202
9 * * * Request timed out.
10 335 ms 364 ms 351 ms rbx-g1-a9.fr.eu [178.33.100.238]
11 353 ms 347 ms 360 ms 37.187.36.131
12 * * * Request timed out.
13 369 ms 363 ms 398 ms mwomercs.com [192.99.109.192]

Trace complete.

C:\Users\Luke>

I'm in South Korea and getting routed through the same ****** French server. So now tell me again how its an AU ISP thats causing the issue. I had a 220ms ping before the move and now average 360ms. That's just unacceptable. Especially if they've moved to a cheaper server after I've payed them a bunch of cash. At the very least PGI should set up a VPN or VPS for us in AU/NZ and East Asia.

Not all AUS ISP's are effected predominantly it's Telstra customers; other ISP customers internode and iinet (whateveritwas) are routing correctly as evidenced by traces from said customers in this very thread.

It is quite possible ISPs in the Asia region as a whole may have this problem; we can only decider the traces we are given.

If this was a PGI or host problem you would expect issues for everyone.

[N0MAD]

VanillaG, on 04 September 2014 - 07:36 AM, said:

Those sites are hosted in a different data center and use a different transit provider, Cogent, than the ones that MWO is using. Looking at all of my traces MWO is being hosted by OVH so you would get different routes because they are on totally different providers.

My guess is that issue that Aussie players are seeing is that some segments of the their ISP network are getting routed to Europe because OVH is a large European based host. Their routers most likely see an OVH IP address and assume Europe so try to route through there.

Now thats how i understand it.
But thats very diferent than saying AU ISPs are forcing their routs thru EU, which would mean that its not the ISP fault or intention that their route goes the long way.
Ego, relax son, theres no bad intention in my posts im not insulting anyone or putting anyone down, im trying to figure out the problem and hopefully find a solution.
But pls if you have figured out what the problem is, i mean we know youre a network engineer and probably a guru at it, pls enlighten us so that people can overcome this problem asap.

[Oderint dum Metuant]

N0MAD, on 04 September 2014 - 08:16 AM, said:

Now thats how i understand it.
But thats very diferent than saying AU ISPs are forcing their routs thru EU, which would mean that its not the ISP fault or intention that their route goes the long way.
Ego, relax son, theres no bad intention in my posts im not insulting anyone or putting anyone down, im trying to figure out the problem and hopefully find a solution.
But pls if you have figured out what the problem is, i mean we know youre a network engineer and probably a guru at it, pls enlighten us so that people can overcome this problem asap.

I think you've already had that answer.

I don't recall anyone saying Telstra is forcing it that route; it could just as well be a hickup but until people point it out to them it won't change.

If other AUS based ISP's can get the route correct so can Telstra.

[Egomane]

N0MAD, on 04 September 2014 - 08:16 AM, said:

Ego, relax son, theres no bad intention in my posts im not insulting anyone or putting anyone down, im trying to figure out the problem and hopefully find a solution.

The solution is to contact your ISP and let them know of the problem. Show them the traceroute results from you and from those unaffected, maybe link them to the various threads in this forum. Ask them to investigate the matter.

And if you arn't past 60 years in age, it is biologically unlikely, that I'm your son! So please don't call me that!

[VanillaG]

N0MAD, on 04 September 2014 - 08:16 AM, said:

Now thats how i understand it.
But thats very diferent than saying AU ISPs are forcing their routs thru EU, which would mean that its not the ISP fault or intention that their route goes the long way.

Forcing is a strong word, but the AU (and other country) ISPs that are having issues are routing the traffic through the EU because THEIR routers think that is the best route to go. The only way to fix the issue is to have the ISP routers fixed/changed/updated to not route through the EU for this type of traffic. As evidenced by other regional ISPs that do not have this routing issue it is limited to certain ISPs and might be particular to certain segments of an ISP's network.

Each ISP is responsible setting up the peering agreements to get to other networks. In my particular case, the change has resulted in a 20ms decrease from 100's to 80's ms ping times. When I look at my trace route, all my traffic goes from my state to Texas on my providers internal network and then is shifted to the OVH network in Texas because my ISP and OVH have a peering site in the same city in Texas.

The peering agreements that your ISP have will ultimately influence who quickly your traffic is routed. This is why people in the same country can see different ping times if they are on different ISPs.

[Enileph]

De-syncs is a major problem.
Funny thing is that problems don't get serious for the first two days of migrating. That makes me think that someone has done something wrong after that.

This is especially true that my performance goes down more after the last patch.

[Oderint dum Metuant]

De syncs are a totally separate issue.
Don't confuse the two

[Iyica de Tylmarande]

Brisbane here with Telstra. Went from 200ish ping to 320-380ping. Pretty much can't use ballistics or PPCs or anything that requires you to lead the target now. Also get desyncs every 5 games or so, where as before I managed to play about 50 or so games with only 1 instance of a desync.

EDIT: My ping for other online games such as Dota 2 and Eve Online remain the same so confirming at least it's not a change with my ISP.

Edited by Iyica de Tylmarande, 04 September 2014 - 07:33 PM.

[ShinVector]

Iyica de Tylmarande, on 04 September 2014 - 04:40 PM, said:

EDIT: My ping for other online games such as Dota 2 and Eve Online remain the same so confirming at least it's not a change with my ISP.

Ok... Let me get this straight for some people.

Good to know that your ping didn't change for other game and that means generally there is no issue that happen at Telstra recently.

Issue here is MWO changed Datacentre including ISPs.
Unfortunately this caused Telstra's routing to MWO at OVH Datacentre to get worst.
Optimally the way to resolve this issue is both MWO's ISP and Telstra to work together to find a better route to each other.
Nothing will happen without the MWO and Telstra users reporting the issue to their respective ISPs.

Edited by ShinVector, 05 September 2014 - 06:33 AM.

[GhostBear64]

I was able to fix this issue by cycling my ADSL2 Modem , its the standard Telstra Thompson Router. Just Turned it off waited five and turned it back on. All good , asked the network engineer at my work and he said some the routers have a DNS cache and they dont update it quickly so you get strange results. rebooting the device makes it clear. But hey I guess you all tried that !!

[Aznox]

VanillaG, on 04 September 2014 - 07:36 AM, said:

Those sites are hosted in a different data center and use a different transit provider, Cogent, than the ones that MWO is using. Looking at all of my traces MWO is being hosted by OVH so you would get different routes because they are on totally different providers. My guess is that issue that Aussie players are seeing is that some segments of the their ISP network are getting routed to Europe because OVH is a large European based host. Their routers most likely see an OVH IP address and assume Europe so try to route through there.

This more likely the answer (abeit just a bit more complicated than that).
In BGP (the internet routing protocol) you configure your routers to tell the world :

" I am here*, i belong this** network, and you can join all thoses xxx IP address through me !"

Then all the other routers propagate the information around the world.

As it was said, OVH is a European based hoster (and only arrived in Canada a few years ago) , so most of their IP address are are announced by their routers in France.

One possible scenario, is that OVH has a very big IP bloc announced in France, and recently started announcing a smaller part of that bloc in Canada. This scenario is perfectly supported by BGP, the most "precise" announcement get priority other the larger one.

One "common" problem is that some routers ignore announcement that are "too precise" (old routers that couldn't handle all the internet routing paths otherwise).

So to simplify :

Ovh announce ip 1-256 in Europe
Ovh announce ip 200-256 in Canada

Tesltra/some-other-ISP routers ignore the 56 ip announcement because their routers are limited in the number of internet routes they can keep in memory.

PGI gets ip 222 for their server.

Tesltra think its located in France and route the packets to Europe.
Packets arrive at OVH France, who then relay them to OVH Canada.

I would bet a few mechs on this scenario or a very similar one.

[Skynet2000]

This issue also affects players living in other Asian countries. My ping jumped from 260 to 400+ since last patch. I am using one of the best ISP providers in Hong Kong.

[kiriage]

Was having crap ping..now returned to a normal 270 ..am with telstra..happy days..anyone else noticed a change?