PSE1 SG server Gone?

Drakuloth · December 11, 2020, 10:40pm

Hey all,

I haven’t provided updates on this because I haven’t heard anything back yet, but we just bumped the thread on this to see if there’s any more information. I’ll let you know if we find anything out or need any more information, but the last thing I heard was there was a routing issue being investigated near Sydney. If you’d like to help for the time being, please post WinMTR tests. We can compare these to the old ones and see if the same routing issue still exists.

To do this, let’s run a WinMTR test which catches a latency issue. You want to restart the test every match until you catch one of these problems. Once you’ve caught one red handed, run the test for about 5 more minutes. This will let us look for problems between your house and our servers which may cause this.

If you have problems with posting it due to a link error, go ahead and copy paste this into your next post, and replace “WinMTR goes here” with your test.

~~~~
WinMTR goes here
~~~~

Siouxsie-11261 · December 12, 2020, 5:31pm

WinMTR test to 35.240.194.184 (GSG1) taken from custom game with “Singapore 2” data center selected

|------------------------------------------------------------------------------------------|

| WinMTR statistics |

|------------------------------------------------|------|------|------|------|------|------|

| router.asus. com - 0 | 254 | 254 | 0 | 10 | 21 | 16 |

| lo40.lns09.mlganxd.wa.vocus.network - 0 | 254 | 254 | 6 | 20 | 27 | 23 |

| ae2-14.per01.mlganxd.wa.vocus.network - 0 | 254 | 254 | 7 | 21 | 56 | 38 |

| ae13.cor01.sydnmtc.nsw.vocus.network - 0 | 254 | 254 | 51 | 66 | 143 | 75 |

| as15169.per03.sydnmtc.nsw.vocus.network - 0 | 254 | 254 | 52 | 69 | 210 | 53 |

| 108.170.247.50 - 0 | 254 | 254 | 53 | 66 | 84 | 69 |

| 142.250.61.2 - 0 | 254 | 254 | 144 | 155 | 166 | 160 |

| 108.170.225.101 - 0 | 254 | 254 | 143 | 157 | 206 | 143 |

| 172.253.68.179 - 0 | 254 | 254 | 144 | 158 | 175 | 159 |

| No response from host - 100 | 52 | 0 | 0 | 0 | 0 | 0 |

|____________|||||||

WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider

CunplusJimmy-1821 · December 15, 2020, 1:02am

|------------------------------------------------------------------------------------------|
| WinMTR statistics |

Host - %	Sent	Recv	Best	Avrg	Wrst	Last
RT-AX56U-7578 - 0	761	761	0	0	10	0
203-121-197-1.r.fuzenet.com.au - 0	761	761	1	2	21	2
10.3.2.253 - 0	761	761	1	4	72	2
au-sa-1837-ipe-01-et4-20020008.tpgi.com.au - 1	757	756	1	2	34	3
au-sa-1837-ipg-01-et2-2003.tpgi,com.au - 0	761	761	2	3	21	3
adl-sot-way-crt1-gi5-1-0.tpgi.com.au - 1	757	756	2	2	16	2
41.20.87.203.static.sa.chariot.net.au - 1	757	756	2	2	31	3
adl-adc-lon-wgw1-be-100.tpgi.com.au - 1	757	756	2	2	19	2
be300.akingbrdr11.aapt.net.au - 0	761	761	2	3	21	3
Bundle-Ether12.way-edge902.adelaide,telstra,net - 0	761	761	2	4	21	4
bundle-ether10.way-core10.adelaide.telstra,net - 0	761	761	3	4	21	5
bundle-ether16.exi-core10.melbourne.telstra,net - 0	761	761	16	17	32	17
bundle-ether12.chw-core10.sydney.telstra,net - 0	761	761	26	28	40	27
bundle-ether1.oxf-gw11.sydney.telstra,net - 0	761	761	27	29	388	29
bundle-ether2.oxf-gw10.sydney.telstra,net - 0	761	761	27	28	39	28
203.50.13.94 - 0	761	761	27	28	39	29
i-91.sydo10.telstraglobal,net - 0	761	761	26	27	46	27
unknown.telstraglobal,net - 0	761	761	28	32	91	29
et-0-0-48-br01-eqsy4.as57976,net - 0	761	761	27	43	334	40
et-0-0-0-pe03-eqsy4.as57976,net - 0	761	761	146	149	220	147
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
No response from host - 100	153	0	0	0	0	0
________________________________________________	______	______	______	______	______	______

WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider

Kenji-12929 · December 17, 2020, 5:23am

I’m in Perth, and the server change has made the game unplayable for anyone in my region. The latency is now 130-140ms, and that’s just not good enough for competitive.

The game is going to lose a whole lot of active players from this. I have been a daily player since launch but finally, have given up. I won’t be getting OW2 either since the issue seems likely to carry over to that game.

Also, paying for a VPN is not a fix. The consumers shouldn’t be paying for the mess they made and can afford to fix.

CunplusJimmy-1821 · December 18, 2020, 2:44am

Hi Drakuloth.

Do we have any update on the situation?

yedrellow-1214 · December 18, 2020, 4:35am

VPNs are also not a fix because if you vpn to Singapore from Perth, you will still get comp games in Sydney, granted they will be less often.

Players in Perth need to have good ping to both Sydney and Singapore. By the time we get home from our 9-5s, Sydney is already asleep so we have to play in Singapore. We essentially miss the australian eastern peak due to timezones, and are instead aligned with the Singapore / South Korea peaks. VPNing in to Singapore on the other hand kills our playerpool earlier in the day.

It kills our playable population to be vpning. Perth needs to have good ping to gsg1 and SYD2 or it will kill our city’s access to the game.

Meliodas-12364 · December 20, 2020, 11:46am

Now the servers of GSG1 are giving me 160ms ping instead of usual 120-130

Siouxsie-11261 · December 20, 2020, 4:27pm

the email address of Google’s Singapore data centre for questions and feedback is singapore@google.com so let them know the problem since blizzard isnt helping

yedrellow-1214 · December 21, 2020, 5:48am

I will email them, but I honestly doubt any response at all from google considering I am not their direct customer.

Nicole-1893 · December 21, 2020, 6:02am

Looking over the tests I think the issue is the routing peer between the ISP and the data center, not the data center itself (a data center does not decide how you reach it).

Many of the ISPs in that region are using Telstra for their routing, and it doesn’t seem to be performing as expected (no one wants 160ms ping) to reach the new server location. In the most recent WinMTR above, check out how many Telstra hops the connection is being sent through in the middle of the test – it’s NINE hops. Most WinMTRs that I analyze on this forum contain fewer routing hops than that.

An ISP is the direct customer of the routing peer (Telstra in some cases, but might be different for other ISPs). To get started on a fix, I recommend encouraging the ISP to look into this. Blizzard is the customer of the data center only, not the ISPs or routing peers. While Blizzard can collect data and pass it on from this thread, the ISP and routing peers may not have a direct obligation to work with them (Blizzard) on a resolution.

yedrellow-1214 · December 21, 2020, 6:29am

I mean, the problem is that every isp in australia is having this issue, it’s not a matter of 1 or 2 messing up, it’s literally all of them. Not just telstra, but tpg, Aussie Broadband, iinet and even a person from India earlier in the thread had an analogous issue.

My isp (Aussie Broadband) has tried to contact google already, but has received no response so far.

The crux of the issue is that Google has a datacenter in Sydney, my isp can only control the forward path by manually entering the ips and forcing the forward path to go Perth to Singapore. However even when they do that, they still run in to the issue that Google is forcing all australia routes to go back to the Sydney google POI.

This is causing the pattern for all of the Australian routes where it goes:

Perth ---->Sydney ---->singapore ----> Sydney —> Perth.

My isp literally has no control over the return path, and no isps do in Australia.

I am sure you can see the problem here; you Blizzard, say the problem is between google and their peers and to contact my isp. My isp, which I have already contacted say that the problem is on google’s end and can only manually force the forward path for known ip ranges, and google according to your post say that google in singapore is not obligated to fix their routing.

My question is, if their routing is broken to Australia, and their routing is broken in India and Mauritius, and who knows where else, and isps can not fix it without google’s intervention, then why is google even a valid option for a gameserver at all?

They have left their routes broken, with no method of fixing them, to half the region.

Nicole-1893 · December 21, 2020, 7:13am

Please be aware that I do not work at Blizzard. I’m a forum volunteer. Staff have blue and orange text only, with staff titles by their names.

It does seem possible that Google has a contract with a routing peer that is having issues and they (Google) would be responsible for working that out.

However, this issue doesn’t seem to be highly-reported one (only a handful of threads), and some players probably reported via ticket, but that leads me to believe that not all players in the region are experiencing a high ping. That’s probably why Blizzard asked for the network tests to help narrow it down and forward the info.

My post was meant to offer sympathy since there hasn’t been a response in a while, and also to set some expectations on what Blizzard can and can’t do (an opinion based on 3 years of reading their responses in the forum).

Edit: Fixed a typo

yedrellow-1214 · December 21, 2020, 7:29am

Alright, that’s understandable if you’re not staff. I would point out though that it is the entirety of Australia with this issue due to some of the learnings in this and one of the other threads on the topic. The reason that it isn’t (widely) reported is that it is a subset of a subset that would have this issue be a large one.

The first subset is that it would only increase pings heavily in Adelaide and Perth, with minor increases in Melbourne, but not Sydney or Brisbane.

The second subsetting involved is that of the players in Perth and Adelaide, not all of them even like playing on Singapore to begin with (and would rather play in Sydney), and it’s usually people playing lower population modes (eg deathmatch, or high MMR) or after the Eastern states peak (there is a time difference of 3 hours between Western and Eastern Australia).

I think that’s the reason for the low reports; it’s not that the problem isn’t fundamentally non-existent for most in Western/South Australia, it’s just that Western /South Australian players who were dependent on the previously existing PSE1 are a small subset to begin with.

yedrellow-1214 · December 26, 2020, 10:13am

This is getting kind of ridiculous, I dunno how google have managed it, but they have managed to make their routing even worse. To the extent it is basically unusable.

WINMTR doesnt show it below because it doesnt get through the full routes but:

|------------------------------------------------------------------------------------------|
| WinMTR statistics |

Host - %	Sent	Recv	Best	Avrg	Wrst	Last
DSL-AC3100-F4C0 - 0	85	85	0	0	5	0
loop612451280.bng1.vdc01.per.aussiebb .net - 0	85	85	6	7	18	8
HundredGigE0-0-0-7.core1.vdc01.per.aussiebb .net - 2	81	80	53	53	59	54
be2.core1.nextdc-p1.per.aussiebb .net - 0	84	84	53	53	57	54
HundredGigE0-0-0-19.core1.nextdc-s1.syd.aussiebb .net - 2	81	80	53	53	58	54
be1.core2.nextdc-s1.syd.aussiebb net - 0	84	84	53	53	60	54
119-18-32-167.cust.aussiebb .net - 0	84	84	53	53	58	53
No response from host - 100	16	0	0	0	0	0
142.250.234.212 - 0	84	84	53	53	56	53
108.170.247.82 - 0	84	84	53	54	58	55
142.250.61.2 - 0	84	84	145	146	162	146
74.125.252.211 - 0	84	84	145	146	162	146
172.253.68.253 - 0	84	84	145	145	161	145
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
No response from host - 100	16	0	0	0	0	0
________________________________________________	______	______	______	______	______	______

WinMTR v0.92 GPL V2 by Appnor MSP - Fully Managed Hosting & Cloud Provider

Check this out, hidden behind all those No Response from host messages is this monstrosity:

https:// imgur(dot)com/a/kiDDh1H

230-500 ping, warning symbols, packet loss, everything.

All of this from a server when Amazon ran it had 55 ping and no problems whatsoever. This is getting beyond ridiculous and I am not sure google are even capable of running a gameserver properly.

Meliodas-12364 · December 27, 2020, 12:48pm

Don’t worry - Blizzard will use Google servers because it’s cheaper than Amazon. And the upcoming sequel will sell too. These problems will never be addressed too because we are not the bigger audience here. They won’t lose any money even if countries like Australia, India don’t get to play.

yedrellow-1214 · December 29, 2020, 11:07am

Apparently this problem is also an issue in Sri Lanka and Bangladesh; so that is 4 countries that have broken routing to gsg1

Siouxsie-11261 · December 29, 2020, 3:12pm

pain, nothing but pain

LastRonin-11150 · January 2, 2021, 9:36am

Iv been an Overwatch player from Sri Lanka from launch till now. Ever since they got rid of the Amazon server iv been getting an unstable connection with 120 ms. It used to be 78 ms before they made the change. Blizzard wont do anything about this and will just respond saying that we should contact our ISP’s and configure it from our end. They only care about popular countries with a larger playerbase.

LastRonin-11150 · January 2, 2021, 9:50am

Can someone suggest me the best mudfish nodes? Im trying everything available to lower my ping.

yedrellow-1214 · January 2, 2021, 11:41am

It really depends on where you’re from. If you had okay ping to Singapore before but don’t now, you can always try vpning in to the Singapore Amazon nodes.

Mudfish also can calculate for you which path is the fastest, so that’s the normal method to improve latency.

You can also just leave it on auto with Overwatch singapore as the destination, and it should do an okay job most of the time. Just don’t waste money on subscription vpns, it’s like an order of magnitude more expensive.