Question Random ping spikes, ongoing issue ?

[zeth07]

I've been having ongoing issues with ping spikes and back and forth with my ISP about what's wrong or not and I just want to try and troubleshoot things on my end further cause it honestly makes no sense half the time. The ping spikes are happening rather frequently at least throughout the day when I monitor them. I say frequently mostly as in it's not a one off occurrence most of the time and gets bad the later the day goes on until extremely late.

If I happen to be playing an online game, particularly something that needs very precise internet/frames these types of spikes are noticeable. For example when I saw some bad rollback on SF6 it went from 1f to 4f or more which is noticeably bad and sometimes higher before it goes back down.

I usually notice the bad spikes throughout the hops "randomly" but typically I see it on the 2nd hop and the last hop of course but it's like there's no rhyme or reason to it. For example this one that I took a screenshot of happened at 1:38am Eastern so no one else would be on the internet using up bandwidth in the house and I'm sure it's not exactly a major active time across the neighborhood network either.

But 36.5ms avg up to 81.9ms and sometimes worse is a huge spike relatively speaking. And they happen just random yet "frequently" enough to be annoying for games that care about a stable / non-spiking connection, even on a personal level.

While it hasn't happened on this graph and I haven't seen it throughout the time I've been typing this I've also seen it spike directly on my router/1st hop, while not as severe it would show an avg of maybe 1ms like normal but I'd see spikes of like 9ms or more and sometimes up to like 20ms on the router which doesn't make sense. Typically I am monitoring when no one is home so I really don't know what could be causing it. I know sometimes the packet stuff is that it's waiting to get through like bufferbloat but if basically nothing else is using it I don't see how it could even be held up to begin with.

Just for reference it usually gets way worse in the afternoon / evening so I always suspect it is from oversaturation of the neighborhood since we have like 400MB+ download and no one in the house is remotely close to using that much at any given time even while I'm checking it.

 

[Ralston18]

No need to mask out the hop IP addresses between your router (192.168.1.1) and Google (8.8.8.8).

Likely those IP addresses are beyond your ISP and little can be done with respect to fixing problems there.

However, there may be some pattern to them and/or the location(s) somewhere having storms or other geographical problems.

Make and model modem, router, or modem/router (if combined)?

Yours or ISP provided?

Run "ipconfig /all" (without quotes) via the Command Prompt and post the full results.

You should be able to copy and paste everthing without needed to retype it all.

 

[bill001g]

So you first have to be very careful about how you use tools like this. You have to really understand the limitation of tools like this. They tell lies all the time.
First it is highly unlikely 50 ms difference between max and average is going to be noticeable in most games. Most times you need close to 200ms to really be able to see single packet spikes and even then you have to be very lucky that it happen at just the right time. In most cases you need multiple packets in a row to see a issue.

Your biggest issue is the test data is not very useful. The largest issue with this tool and most other tools is routers are designed to favor passing real traffic over responding to ping/test data. So if the router has a choice of responding to a ping packet or passing a packet to the next router in the path it will chose to pass the data and delay responding to your ping. Since your game traffic is not ping packets this test is not really valid because the router itself is not responding to it.

If you see packet loss or very large delays it can be more significant but small delays like you are showing can easily be testing error.

Next your data is even more suspect. If the problem was actually in hop 2 every hop past it would show the same issue or actually be slightly more. In this case the delta is 20ms for hop 2 (which is extremely small) and the hops past it are closer to 10ms delta. If there was some issue with hop 2 it would have this 20ms in every hop past it.
Hard to read much into this though because the numbers are so small. if this was 100ms or 200ms it would be much much more obvious.

The delay in the final hop means very little. It could be network but it could also be the server.

I strongly suspect you decided the network was the problem, found a tool that showed some small abnormal data and are now trying to confirm your original theory. Unfortunately this tool is not really showing the problem. It does not mean there is no network issue it is this tool just does not show it well enough to confirm it.

Bottom line is you can do nothing to fix anything outside your house. Even if this was real what it generally means is there is other people traffic competing for bandwidth and your traffic is being delayed. We have to wait until the greedy ISP decide to sell premium internet connection where you have pay more than other people...kinda like toll lanes on the highway.
This to a point exists with some of the very special vpn that have direct access to some of the more popular games but it only fixes very narrow issues. It is more for say someone in asia whose ISP does not lease space on the most direct undersea fiber routers. In the USA and EU these VPN services tend to not do much even when there are actual problems.

 

[zeth07]

So you first have to be very careful about how you use tools like this. You have to really understand the limitation of tools like this. They tell lies all the time.
First it is highly unlikely 50 ms difference between max and average is going to be noticeable in most games. Most times you need close to 200ms to really be able to see single packet spikes and even then you have to be very lucky that it happen at just the right time. In most cases you need multiple packets in a row to see a issue.

Your biggest issue is the test data is not very useful. The largest issue with this tool and most other tools is routers are designed to favor passing real traffic over responding to ping/test data. So if the router has a choice of responding to a ping packet or passing a packet to the next router in the path it will chose to pass the data and delay responding to your ping. Since your game traffic is not ping packets this test is not really valid because the router itself is not responding to it.

If you see packet loss or very large delays it can be more significant but small delays like you are showing can easily be testing error.

Next your data is even more suspect. If the problem was actually in hop 2 every hop past it would show the same issue or actually be slightly more. In this case the delta is 20ms for hop 2 (which is extremely small) and the hops past it are closer to 10ms delta. If there was some issue with hop 2 it would have this 20ms in every hop past it.
Hard to read much into this though because the numbers are so small. if this was 100ms or 200ms it would be much much more obvious.

The delay in the final hop means very little. It could be network but it could also be the server.

I strongly suspect you decided the network was the problem, found a tool that showed some small abnormal data and are now trying to confirm your original theory. Unfortunately this tool is not really showing the problem. It does not mean there is no network issue it is this tool just does not show it well enough to confirm it.

Bottom line is you can do nothing to fix anything outside your house. Even if this was real what it generally means is there is other people traffic competing for bandwidth and your traffic is being delayed. We have to wait until the greedy ISP decide to sell premium internet connection where you have pay more than other people...kinda like toll lanes on the highway.
This to a point exists with some of the very special vpn that have direct access to some of the more popular games but it only fixes very narrow issues. It is more for say someone in asia whose ISP does not lease space on the most direct undersea fiber routers. In the USA and EU these VPN services tend to not do much even when there are actual problems.

Is there anything else I could do to test out the problems?

I don't think it is impossible to not be a problem on my own end it's just that everything else that I've tried or the ISP has "fixed" makes it seem like it isn't a problem within our network/lines anymore.

It was much worse before but they've changed the lines coming to the house and we've tested other modems and also without the router, but since it gets worse during certain times it makes it seem like congestion related but I don't know how that relates to the "spikes".

While you say a 50ms difference wouldn't be noticeable in a game, that is still almost 3 frames of "lag", which would be obvious in something like a fighting game because it has a hitch to it with how rollback frames work, which is why I used that as an example. If something only has 1f of rollback it is fine, but if it **suddenly** goes up to 4f of rollback it looks really bad and very apparent.

The reason I've been using pingplotter is to at least try and see if I can notice the spikes in-game while they also show on pingplotter. While I know Google's DNS isn't the same as a game server (or in SF6 P2P) if the spike is happening bad enough in my local area but after my house I have to imagine it would still impact anything I'm doing and be visible in-game like I've seen.

I don't know if it is related to anything but on my modem I also keep seeing these repeated messages in my log, they at least aren't T3/T4 timeouts so maybe they are nothing and can be ignored but I feel like it seems abnormal.

It's just alternating like this for presumably the entire allowable length of the log:

- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13.
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 11 13; New Profile: 9 13
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 11 13; New Profile: 10 13.
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 10 13; New Profile: 11 13

I removed some mac address stuff but otherwise it seems like these loop forever.

 

[Ralston18]

Repost the Pingplotter results without masking the IP addresses.

Especially between hops 2 - 7 (inclusive).

2 and 7 appear to be the same URL....

Make and model info?

"ipconfig /all" results?

 

[zeth07]

Repost the Pingplotter results without masking the IP addresses.

Especially between hops 2 - 7 (inclusive).

2 and 7 appear to be the same URL....

Make and model info?

"ipconfig /all" results?

What's the relevancy of doing ipconfig/all for this case? I don't see anything abnormal there.

For the IP addresses:
- 2nd hop is my ISP.
- 3-6 hops are all different 10.xxx.xx.xx but come up as "private" ones.
- 7th hop is another from my ISP but different.
- Hop 8,9,10 are Google.

Modem: Netgear Nighthawk CM2000
Router: Netgear Nighthawk XR1000v2
ISP: Armstrong Cable
Wired connection on anything that matters.

 

[bill001g]

"While you say a 50ms difference wouldn't be noticeable in a game, that is still almost 3 frames of "lag", which would be obvious in something like a fighting game because it has a hitch to it with how rollback frames work, which is why I used that as an example. If something only has 1f of rollback it is fine, but if it **suddenly** goes up to 4f of rollback it looks really bad and very apparent."

The video itself does not flow over the internet connection. There actually is very little direct relationship between the data being sent between the server and your client and what is being displayed. What is happening is the server is sending its estimation of the status of the server at the time it predicts you are going to receive the data. The rate it send these update varies a lot between games but it is not even close to the frame rate. You can actually completely lose some of these updates and it will make no difference. These games would be unplayable if every time you get tiny delays the client would be desynced from the server. The client tends to be very good at guessing what is going to happen next.

"- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13.
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 11 13; New Profile: 9 13
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13"

These are bit more concerning. In extreme cases this can cause you modem to drop and retrain. They indicate some issue with the configuration I think but I have not hands on done cable modem stuff on the ISP side for a huge number years and stuff has changed. This is something the ISP would have to fix you can do nothing about it since you have no option to change anything about how the modem is configured.

This though I don't think causes the issue you are looking at. This would cause packet loss or longer outages. Delays are always caused by data being held in a buffer. Cable connections only hold/delay data if there is a overloaded connection. Unlike say wifi they do not re transmit data that has errors they just discard it and you see packet loss.

But lets assume everything you say is true what do you think you can do about it. The ISP does not even really promise your bandwidth they always use "upto". They will for sure not promise any latency....especially outside their network. The connection between your house and the ISP first node is shared by many people. If a number of people sharing your cable would use a lot of bandwidth at the same time you will see small spikes in the latency. The ISP can not really fix this....they only make money by telling lies that everyone can use 1gbit where their total bandwidth is maybe 3gbps.

The ISP will fix packet loss ...in the second hop..but they will do almost nothing about other issues. You have already tried replacing the modem, this too is unlikely but is soemthing to try.

 

[zeth07]

"While you say a 50ms difference wouldn't be noticeable in a game, that is still almost 3 frames of "lag", which would be obvious in something like a fighting game because it has a hitch to it with how rollback frames work, which is why I used that as an example. If something only has 1f of rollback it is fine, but if it **suddenly** goes up to 4f of rollback it looks really bad and very apparent."

The video itself does not flow over the internet connection. There actually is very little direct relationship between the data being sent between the server and your client and what is being displayed. What is happening is the server is sending its estimation of the status of the server at the time it predicts you are going to receive the data. The rate it send these update varies a lot between games but it is not even close to the frame rate. You can actually completely lose some of these updates and it will make no difference. These games would be unplayable if every time you get tiny delays the client would be desynced from the server. The client tends to be very good at guessing what is going to happen next.

"- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13.
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 11 13; New Profile: 9 13
- DBC-REQ Mismatch Between Calculated Value for P1.6hi Compared to CCAP Provided Value
- US profile assignment change. US Chan ID: 8; Previous Profile: 9 13; New Profile: 11 13"

These are bit more concerning. In extreme cases this can cause you modem to drop and retrain. They indicate some issue with the configuration I think but I have not hands on done cable modem stuff on the ISP side for a huge number years and stuff has changed. This is something the ISP would have to fix you can do nothing about it since you have no option to change anything about how the modem is configured.

This though I don't think causes the issue you are looking at. This would cause packet loss or longer outages. Delays are always caused by data being held in a buffer. Cable connections only hold/delay data if there is a overloaded connection. Unlike say wifi they do not re transmit data that has errors they just discard it and you see packet loss.

But lets assume everything you say is true what do you think you can do about it. The ISP does not even really promise your bandwidth they always use "upto". They will for sure not promise any latency....especially outside their network. The connection between your house and the ISP first node is shared by many people. If a number of people sharing your cable would use a lot of bandwidth at the same time you will see small spikes in the latency. The ISP can not really fix this....they only make money by telling lies that everyone can use 1gbit where their total bandwidth is maybe 3gbps.

The ISP will fix packet loss ...in the second hop..but they will do almost nothing about other issues. You have already tried replacing the modem, this too is unlikely but is soemthing to try.

I understand what you're getting at, but the main thing at least from my perspective is if I am seeing "problems" at the 2nd hop and we've tried nearly everything within the home network then it feels like something I could bother my ISP about to look into and fix whether it's equipment related or lines beyond my house.

When I call the techs they keep seeing the ping spikes on their end and sometimes they test it and say they do notice it with neighbors but other times they don't. I've got it escalated enough that they keep mentioning them telling the local offices to look into it or some other tech person higher up to better understand it.

The last time I called I got disconnected from one guy cause he tried to reset my modem which we need for an LTE Extender cause our location gets a bad signal, but when I called back I got lucky enough to get a guy that was I guess deeper into the tech level side and he noticed the problems I've been talking about. He also let me send him that modem log stuff but he also said he would send it along to someone else who could understand even more than him.

I haven't heard back from them since it was back on Friday and would have to wait til Monday or later but I did see one of the guys sitting outside in their van so maybe they were messing with stuff but I still see problems so who knows.

I know they only list the speeds as "Typical" but at least in our case their FCC Broadband Consumer label does in fact also list Latency, but it still uses the phrasing of "Typical Latency 17ms". I'm sure that implies to their server but if I'm seeing spikes well beyond 17ms to their server even within a reasonable margin of error that surely means there's a problem.

But again it seems like an impossible solution to find on my end which is why I'm trying to look into whatever I can across the internet.

 

[bill001g]

They are doing very good their "average" latency is 17.2 ms on their ping plotter.

I really wish games had some ability to show what is actually causing the lag. What you see on this forum all the time is people looking at the so called "ping" time in a game. What happens is the game sends it ping and the response gets placed in a buffer. The game gets busy say rendering a frame and then after its done it looks at the buffer. It blames all the delays on the network and pretends the rendering delay did not exist.

What is extremely strange is people fix so called network lag issues by adjusting the video settings. This should make no difference at all since the video itself never flows over the network.

Maybe try to set the option to much lower quality and maybe even lower the resolution. If there is a actual network issue it really shouldn't matter what setting you have the video at you should still have a issue.

 

[Ralston18]

"ipconfig /all" is a commonly used network troubleshooting tool. And you are correct that nothing seems abnormal.

It does not show if you have a wireless adapter installed or not. If so, the wireless network adapter should be disabled. There should only be one enabled network adapter.

However, you masked out the IPv4 Preferred Address which is just the DHCP IP address that the router has provided to the computer.
And if the masked octet is in the range of allowed DHCP IP address - no problem. My router is also 192.168.1.1 and my computer is currently using 192.168.1.118 which is within the range and number of DHCP IP addresses allowed to my router.

There is no harm in showing mac - what you do not want to reveal is the IP address that your ISP has provided to the modem. You can use "What is my IP Address" to discover that IP.

Two configurations changes to consider: Disable IPv6 and set the DNS servers to Google at 8.8.8.8 and 8.8.4.4. Very much a routine "fix" that can prove helpful in some situations.

= = = =

What I truly do not understand are those four 10. IP addresses. Those IP addresses are all within one of the three private IP address ranges used by thousands of small home and business networks.

FYI:

https://www.lifewire.com/what-is-a-private-ip-address-2625970

Specifically 10.0.0.0 to 10.255.255.255

Your router 192.168.1.1. is connected to something using 24.101.102.1 indicated as being your ISP (shown as being dynamic.acs etc.).

Then the path goes through those four private IP addresses and then again to the ISP. Not at all what I would expect.

Which could be a network loop of some sort. And may be related to or indicated by the logs shown in Post #4 and #7. That alternating or toggling between profiles.

After that then out into the internet world to Google.

There are clearly problems with the presence of those private devices. If anything perhaps some configuration issue or problem(s) with the modem....

Ask your ISP to check out or otherwise explain them. If possible....

Hopefully, someone else here will be able to comment as appropriate. Or correct some error of commission or omission on my part.

 

[zeth07]

"ipconfig /all" is a commonly used network troubleshooting tool. And you are correct that nothing seems abnormal.

It does not show if you have a wireless adapter installed or not. If so, the wireless network adapter should be disabled. There should only be one enabled network adapter.

However, you masked out the IPv4 Preferred Address which is just the DHCP IP address that the router has provided to the computer.
And if the masked octet is in the range of allowed DHCP IP address - no problem. My router is also 192.168.1.1 and my computer is currently using 192.168.1.118 which is within the range and number of DHCP IP addresses allowed to my router.

There is no harm in showing mac - what you do not want to reveal is the IP address that your ISP has provided to the modem. You can use "What is my IP Address" to discover that IP.

Two configurations changes to consider: Disable IPv6 and set the DNS servers to Google at 8.8.8.8 and 8.8.4.4. Very much a routine "fix" that can prove helpful in some situations.

= = = =

What I truly do not understand are those four 10. IP addresses. Those IP addresses are all within one of the three private IP address ranges used by thousands of small home and business networks.

FYI:

https://www.lifewire.com/what-is-a-private-ip-address-2625970

Specifically 10.0.0.0 to 10.255.255.255

Your router 192.168.1.1. is connected to something using 24.101.102.1 indicated as being your ISP (shown as being dynamic.acs etc.).

Then the path goes through those four private IP addresses and then again to the ISP. Not at all what I would expect.

Which could be a network loop of some sort. And may be related to or indicated by the logs shown in Post #4 and #7. That alternating or toggling between profiles.

After that then out into the internet world to Google.

There are clearly problems with the presence of those private devices. If anything perhaps some configuration issue or problem(s) with the modem....

Ask your ISP to check out or otherwise explain them. If possible....

Hopefully, someone else here will be able to comment as appropriate. Or correct some error of commission or omission on my part.

I think on my router itself ipv6 is already disabled but I also did it on my ethernet on the PC. I have my wifi stuff disabled on my pc just so it wouldn't even have that as an option.

For DNS I ran that DNS Speed Benchmark that I saw recommended elsewhere and think I assigned whatever was fastest for those. However on my PS5 I'm pretty sure I specifically assigned google or cloudflare.

I feel like I asked my ISP about the other IPs previously and they acted like they had no clue what I was talking about, which doesn't make sense when it has to go through two of their nodes inbetween anyway. I'll probably have to just poke at them again about it I guess.

 

[bill001g]

What I truly do not understand are those four 10. IP addresses. Those IP addresses are all within one of the three private IP address ranges used by thousands of small home and business networks.

This is very common. No end user is actually going to try to log into the router itself so it really doesn't matter what IP they use. If they happen to duplicate some IP all that happens is you get very strange results because some local device in your network might respond with a corresponding very low latency. This is also the cause of * * * entries in a tracert because some ISP or even your own router might filter private ip addresses.

The actual data stream itself does not have that IP as either the source or destination IP so it will not actually be sent to the router itself. This router will purely use these IP to decide what link to place the data on. Again one of strange things that can happen when you try to use trace/ping to test data.

The ISP do this so they don't have to waste actual IP for device. These IP really are for the ISP to access its routers and possibly test things inside their network. Making this even more messy is the ISP can actually configure the router to give any IP back in response to trace so you really never know if it is giving you the IP of the router itself or one of the interfaces on the router.

They of course can configure it to not respond at all.

 

[Ralston18]

@bill001g

Thank you.

So if I understand correctly what is happening it is that @zeth07 's Ping plotter tests are going through ISP devices set up by the ISP for internal test purposes and configured to specifically use private IP addresses.

Messy indeed.

= = = =

In any case, end users can only affect/control what is in their own network up to and including modem, router, or modem/router.

The ISP controls what what is between the customer's devices and and the ISP's devices.

Beyond the ISP's devices - the ISP can do little.

If that is indeed the operating environment I truly do not know what else to suggest.

 

[dengamle]

Regarding the "US profile assignment change" events. The modem is being told to use a different modulation profile. It's a "switch", so there is no "training time".

It happens to avoid packet loss, but your ICMP measurements are loss-free, so I don't see a direct correlation.

 

[Ralston18]

"The modem is being told to use a different modulation profile."....

Where does that instruction come from? Why/when etc...?

 

[dengamle]

"The modem is being told to use a different modulation profile."....

Where does that instruction come from? Why/when etc...?

The command comes from the CMTS and is triggered by some noise event. When the noise clears up, the modem can go back to a higher modulation profile.

 

[Ralston18]

Thank you.

Nothing that the end user / subscriber can affect other than reducing or stopping noise events (if at all possible) on the modem end.

All very much up to the ISP - correct?

 

[zeth07]

I just wanted to provide another example since I feel like the ones provided maybe didn't paint a good enough picture of the type of stuff I keep seeing.

In terms of our own network bandwidth usage I believe they upgraded it to 500MB recently and divided by 8 would 62.5 Mb per second if I'm understanding it right as best possible case scenario.

Normally whenever I start monitoring the network I would see my brother AT BEST using 6 Mb a second, which is well below our max. If at the upper limit of each family member doing 6Mb a second we shouldn't be coming close to 500MB / 62.5 Mb speeds.

Then I went to check my modem log to see if any error messages correlated to the timing and while nothing shows for this specific time I am seeing another that says:

Around Thu Apr 10 22:15:31 2025

"RNG-RSP CCAP Commanded Power Exceeds Value Corresponding to the Top of the DRW"

and "Dynamic Range Window violation"

"Received Response to Broadcast Maintenance Request, But no Unicast Maintenance opportunities received - T4 time out"

"No Ranging Response received - T3 time-out"

"No Maintenance Broadcasts for Ranging opportunities received - T2 time-out"

Then it goes back into the other error messages I posted before.

 

[zeth07]

Now this is happening when I checked a speed test...I even did a power cycle of the modem and router just to be sure it wasn't some random quirk and it still showed up like this. We're listed as 500/20 and usually I don't see actual speed issues but now it's doing this as well.

 

[bill001g]

Your latest information is a very different issue. This the ISP will likely be able to fix. This is actually a very good example of using ping plotter to show a issue.

It show the issue starting in hop 2 and continuing all the way to the end.

Packet loss is caused by a number of things but in many cases it is something very simple like a dirty connection or some other cable damage outside you house. The ISP should be able to locate it with their equipment.

The messages in the modem are tough to say sometimes but things like no ranging response. time out is the pretty much the same thing as a ping command saying gets no loss.

Although it is really the ISP job you can look at the signal levels and there will likely be a screen that show the number of packets that have correctable and uncorrectable errors.