Let me start by saying I’m a network admin, so if you need any additional information just ask away.
For a few weeks now I’ve had absolutely abysmal latency every 30 or so minutes to the game, to the point that the game halts completely and only comes back when the lag ends, which usually lasts anywhere from 1 to 10 minutes.
The problem would eventually stop happening, until it came back a few days later.
Lately it’s been happening nonstop every 10 to 60 minutes, to the point the game has become completely unplayable.
I’m connecting to US East/Central and it seems the issue only happens for these regions, as I’m monitoring US West and Oceania and I don’t see latency spikes on those.
Here’s what it looks like during the issue.
i.imgur . com/zCUKDLw.png
Here’s what it should look like.
i.imgur . com/XpsYoje.png
Here’s how the latency spikes look on a graph.
i.imgur . com/wpxa7wp.png
This is for the last 7 days.
i.imgur . com/7vNer8i.png
For what it is worth, I’m on the east coast and I play on a central server. I have the same issue, mine for example is normally Home 32ms, World 32ms and nearly never a DC. Today it has been Home 32ms, World 800-4000 and DCs every few seconds. Wow is completely unplayable at the moment. (Side note, today is my first day playing in a week due to vacation, so I can’t say about the last 7 days prior to today.)
Myself and several others have done traces, and it seems to be a Level3 Node between us and one of the data centers that is causing the issue. Using a VPN to create a route that bypasses that node resolves the issues.
Your test shows exactly where the problem is happening: the Equinix Miami peering node. As a network admin, you probably know that the node is not owned by your ISP or Blizzard. Both companies can alert them of the problem, but the fix and timeline are up to Equinix.
If there are a lot of players experiencing this, Blizzard network engineers already know about it since they monitor all incoming pings and they would probably reach out if it continues. However, you should let the ISP know, too.
Beyond that, using a VPN to route around the Equinix Miami node will likely resolve the issue for the time being.
The latency jump is at 137.221.76.33.
I’m not sure how you can conclusive know this is a PIX, at least their peeringdb doesn’t list it, but either way this router is owned an operated by Blizzard, meaning it’s up to them to fix this connection.
We can check this both by checking whois of the IP and of AS57976 which is in the PTR record in the mtr.
The only 2 parties directly responsible for this connection are Blizzard and Sparkle, and since I am a paying Blizzard customer but not Sparkle the obvious choice is coming for them.
Not confused at all. The first latency jump of 100ms is between 213.144.160.100 and 89.221.41.175—which has nothing to do with Blizzard. The second large latency jump is at this host: ae1-br01-eqmi2.as.net. The ping jumps from 102ms to 356ms and continues to climb between Equinix nodes. Equinix is the party responsible for ae1-br01-eqmi2.as.net.
The final IP in the test is the Blizzard server, but you accumulated the high ping before reaching it.
Please check the other mtr which shows how it should be when not having the lag spikes.
As mentioned in the title, I am playing from Brazil, typical RTT is ~140ms.
The 100ms jump you’re pointing out is normal, as we can see from the nominal mtr, the Blizzard one should be at 140ms instead of over 300ms with packet losses.
Equinix is the party responsible for ae1-br01-eqmi2.as57976 .net.
The IP is Blizzard’s and the AS in the PTR record is also Blizzard’s.
This might be Equinix’s IDC, but the router is Blizzard’s.
Again, I work in this field, this is Blizzard’s responsibility, not the IDC owner/operator (even if the issue is in a fiber which Equinix is responsible for, it’s still Blizzard’s responsibility to open a support ticket with Equinix).
Since the forums are for players to help each other, you’ll need to open a ticket with the staff to inform them of what you believe are their responsibilities.
Expliquei isso antes, isso não é uma “empresa”, isso é a Blizzard.
Bota no google “whois”, isto é uma ferramenta para identificar informações de domínios, IPs e Autonomous Systems (identidade necessária para se conectar na internet com seus IPs).
A Equinix que você vê na entrada PTR relativa a este IP refere-se a localidade deste roteador, o importante mesmo é o IP em si e até mesmo o que está no lado direito da entrada PTR, as57976.
Dando um whois no IP, podemos ver que o IP é da própria Blizzard.
Dando um whois no AS, podemos ver que é o AS da Blizzard.
Ou seja, isto é um roteador da própria Blizzard cujo eles alugaram espaço no datacenter da Equinix pra colocar lá e operar de lá.
Isto é um problema de responsabilidade direta da Blizzard. Mesmo que a falha técnica seja algo no datacenter da Equinix (por exemplo, a fibra que eles passam está ruim), é de responsabilidade da Blizzard abrir um ticket de suporte com eles, uma vez que eles são clientes da Equinix.
I’ve been having absurd lags since the last month.
my internet is perfectly normal, netflix, stream, google. My MS has always been 132 and suddenly this game has a huge lag (5, 14 seconds of lag) and several times I am summarily disconnected from the game.
It’s obvious that the problem is with Microsoft, how could it just be that the game is lagging and everything else is working normally and the problem is not with the game. Microsoft has been scrapping and cheapening the service since acquiring Blizzard to increase profit.
If they don’t hire a cleaning lady to clean these cables and buy better equipment I’m going to have to cancel the sub because wow is unplayable.
Extract the contents of the .zip file to your desktop.
Open the WinMTR folder, then open the folder that matches your version of Windows (32-bit or 64-bit) and run WinMTR.exe.
Enter the IP address you want to test in the Host field. Select a game from the dropdown below for a list of IP addresses.
5.Note: If multiple IPs are listed, run a WinMTR test for each IP address.*
Click Start, then launch the game you’re testing. If no connection issue is found in the first 5 minutes, restart the test.
When a connection error occurs, play for at least 5 more minutes, then click Stop in WinMTR.
Click “Export Text” and save the .txt file to your desktop. Name this file WinMTR.
Actually completely true. All of the streaming services download 30+ seconds of movie as a buffer. which is plenty of time for the service to resend any lost packets . You can stream movies with no issues with 500+ms and 20 percent packet loss . like i said there is basically no streaming service that is actually affected by packet loss unless its severer. you might see a issue at 25 to 30 percent packet loss. Google sends so little data that it is also really un affected by packet loss unless its sever. The worse you connection the larger the streaming service buffers.
interesting that you backtracked from your original statement of “all of which are unaffected” to “only if it’s 25% or higher”.
also if you actually looked at the MTR i provided in the OP, packet loss is at 50-90%.
and your claim of 25% is untrue as well. all TCP connections are severely slowed down by even around 5% packet losses due to additional round-trips for acks. (again, i work on this field, i have actual expertise here)
They are pertly much un affected the worse you connection the more of the movie that is pre downloaded. Most people have enough bandwidth that even 25 % in reality is not going to affect streaming. Good example i have a 500K connection it would need to be around 90 percent before I would notice if all i did was stream movies. Even with all the packet loss i would still have a 50K connection