Sonic.net

Hi,
I play League of Legends (connected wireless) and have been experiencing abnormal ping spikes to the point of it being unplayable. My typical ping in around 73ms to 600,000ms and experience about 5 times in 40 minutes with a duration of 1-2 minutes each time. Knowing this I ran a network test using netalyzer and stated that I have a network buffering issue and the problem is this:
--
A problem in this test means your network can't "Walk and chew gum" at the same time. You can conduct large file transfers (such as using file sharing programs), or you can perform interactive tasks (like websurfing or VoIP calls) without the buffer being a problem.

The problem occurs if you try to do both types of activity at the same time: if you try to run file transfers concurrent with other activity, the other activity will feel very "slow", as the file transfer fills up the large buffer, adding latency to all other traffic.
--

This is the link to the test results.
http://netalyzr.icsi.berkeley.edu/resto ... -417e-b4a3

What is the solution I can implement?

Try it wired instead of wireless.

Try connecting via wired connection to see if you get the same results.

zhao156 wrote:Hi,
I play League of Legends (connected wireless) and have been experiencing abnormal ping spikes to the point of it being unplayable. My typical ping in around 73ms to 600,000ms and experience about 5 times in 40 minutes with a duration of 1-2 minutes each time. Knowing this I ran a network test using netalyzer and stated that I have a network buffering issue and the problem is this:
--
A problem in this test means your network can't "Walk and chew gum" at the same time. You can conduct large file transfers (such as using file sharing programs), or you can perform interactive tasks (like websurfing or VoIP calls) without the buffer being a problem.

The problem occurs if you try to do both types of activity at the same time: if you try to run file transfers concurrent with other activity, the other activity will feel very "slow", as the file transfer fills up the large buffer, adding latency to all other traffic.
--

This is the link to the test results.
http://netalyzr.icsi.berkeley.edu/resto ... -417e-b4a3

What is the solution I can implement?

I just tried it on the wired connection and there results are

http://n2.netalyzr.icsi.berkeley.edu/re ... d2-a02b/rd

My downlink is amazing at 52ms but my uplink is at 1100ms any advice?
I think thinking of having a lower upstream cap might do the trick?

Nice to see you were able to get improvements. There are still a couple things that could cause this. I think a simple test to futher isolate the problem would be to run the test on the wired connection again and with all other devices disconnected and post back the results along with your network setup.

zhao156 wrote:I just tried it on the wired connection and there results are

http://n2.netalyzr.icsi.berkeley.edu/re ... d2-a02b/rd

My downlink is amazing at 52ms but my uplink is at 1100ms any advice?
I think thinking of having a lower upstream cap might do the trick?

If you happen to already have another router that can be used behind DMZ+/passthru/whatever then it might be worth testing out. This allows the CPU of the primary router to concentrate on the modem work while the backend router deals with routing. It does cost more power draw but it might resolve the current issue.

I like the simplicity of just using one modem/router but the 5031nv that is provided for my Fusion FTTN X1 has lots of problems that are resolved by using DMZ+. I think you have "pure" Fusion but it might be worth looking into.

The problem is caused by bufferbloat. https://en.wikipedia.org/wiki/Bufferbloat

If you have another computer that hasn't been tweaked at all, rerun the test there to see if it's caused by your router. Otherwise, you can look at that Wikipedia entry for ways to mitigate the problem.

I wonder if the issue is being caused by congestion of your upstream connection. This could occur as a result of "cloud sync" type activities, for example iCloud uploads, Dropbox, backups, etc. In theory, those tools should be well behaved about not monopolizing the upstream, but I wonder if there may be some which are not.

dane wrote:I wonder if the issue is being caused by congestion of your upstream connection. This could occur as a result of "cloud sync" type activities, for example iCloud uploads, Dropbox, backups, etc. In theory, those tools should be well behaved about not monopolizing the upstream, but I wonder if there may be some which are not.

Shouldn't that be handled by QoS ?

pockyken007 wrote:

dane wrote:I wonder if the issue is being caused by congestion of your upstream connection. This could occur as a result of "cloud sync" type activities, for example iCloud uploads, Dropbox, backups, etc. In theory, those tools should be well behaved about not monopolizing the upstream, but I wonder if there may be some which are not.

Shouldn't that be handled by QoS ?

I don't believe there's any QoS favoring outbound ACK packets, so saturation of upstream can affect inbound traffic.