How can I determine whether persistent VDSL bufferbloat is caused by my modem/ro...

I am trying to determine whether my latency problems are caused by my ISP or by my VDSL modem/router. Now before you just suggest buying a new expensive thing. That is an investment and I might just spend money to find out it's my ISP all along. However, if it is indeed my ZXHN H168N V3.5 I can save and get better hardware. But I gotta rule that out first if possible.

Now note that I'm in Africa and the following numbers are good for my country.

• 30 Mbps download / 3 Mbps upload package
• Actual speeds are usually 26-28 Mbps down and 2.5-2.8 Mbps up
• Idle latency 8-10 ms. I'm fine with that speed. I just want it to work.
• Download latency typically 80-130 ms.

All the numbers above are what shows when I run a speedtest @Speedtest.net, the only site my ISP accepts. Tens of tests over months btw. This is making gaming a nightmare but more on that later.
On that site I can't see an exact server address or anything. Result only shows the numbers above and gives a city name which is the same as my city.

• Waveform Bufferbloat test consistently grades the connection a good old F.

Setup

• VDSL
• ZXHN H168N V3.5 modem/router combo
• Ethernet from router directly to PC. All tests on PC.
• Tests done at different times.

Since I'm in Africa I naturally get a ton of latency to gaming servers in, where I usually play. But the issue is so far outside the normal range. In several games. Such as: CS, Darktide, Valorant, the finals.

I know that within those games there are different servers and yes some are worse. But the theme in all of them is the same. Consistent measurable latency that I feel I'm 1-2 seconds behind everyone playing. With humans it's worse but even in a horde shooter like darktide the bots get a head start on me.

Troubleshooting already tried

• Ethernet connection only
• Wi-Fi disabled
• Different DNS servers
• IPv6 on and off
• Direct connection without splitter.
• Fresh Windows installation, that was happening anyway not that I did to try it, but I digress.
• ISP technician checked the central hub for my area, not sure what to call it, that houses the phone/internet lines and said everything is fine.
• ISP technician came to my house and checked the line and reported no physical faults in all of my setup.
• Tried looking for smart queue management in my router page but no luck.
• Tried with an LLM, some of the tests below are their requests, but I don't trust them much

While saturating the download connection:

Ping to my router ():

66 packets sent 66 received 0% packet loss Maximum latency 1 ms

Simultaneous ping to 8.8.8.8:

66 packets sent 66 received 0% packet loss Minimum 8 ms Maximum 135 ms Average about 100 ms

An additional problem appeared today and here is what I observed. Please notice this is an isolated event that happened and ended. The above issue is SEPARATE FROM THAT. The above is a consistent problem. This just another problems

The connection became nearly unusable for about 9 hours. From about 6 PM to 3 AM.

• Speed tests still reported normal numbers.

• Downloads still reached expected speeds.

• Web browsing became extremely slow.

• Discord, Reddit and Facebook frequently stalled. Like I'm talking taking a minute to load.

• Online games experienced severe packet loss.

• I tried the usual. A PC restart and a router restart, removed the power.

• And I ran some tests during that time

• Ping to 1.1.1.1 averaged roughly 159 ms with 5% packet loss.

• Ping to 8.8.8.8 showed approximately 61% packet loss.

After those 9 hours the problem solved itself without me making any changes to my setup or anything.

Since this has being going on for months I called my ISP 11 times before. I was always patient and calm. When they suggested something I did it. When the said wait 48 hours I did. But not a single person I talked to even knows what latency is in the first place. And on more than one occasion they tried to gaslight me. I don't want to attack them now but I'm explaining why I'm doing this and asking you here.

That's why I need to arrive at a conclusion on my end before I start trying to escalate this. Either try with them, again, or try to go to whatever government agency that handles this. And, again, if it's that my ZXHN H168N V3.5 is bad I'll save and upgrade. But I need to isolate the issue to take action accordingly.

What can I do to figure out exactly what is the problem?

Even if all the downstream bufferbloat is caused by your ISP, it probably doesn't matter as much as you think; you can still fix it by taking action on your end even if your ISP won't lift a finger to help you.

DSL ISPs are famous for having massive bufferbloat problems and not seeming to know or care anything about it. But luckily, even if the bufferbloat is on the ISP's side, you can still fix it from your side.

To do this, you have to have to have an SQM-enabled router (that is, it needs to have a good SQM algorithm like CAKE or FQ-CoDel enabled). You need this router to be at the head of your network so all traffic between the Internet and anything on your home network goes through this router. Then you need to configure that router's bandwidth-limiting (might be called something like "traffic shaping" or "Quality of Service" / QoS ) to make that router a slight bottleneck in both directions. So if you usually get 30 Mbps downstream, you set its downstream limit to 29.5 Mbps. If you usually get 3 Mbps upstream, you set its upstream limit to 2.5 Mbps. Please note that the values I'm suggesting here are for illustration purposes only; there are online guides that might be able to give you more precise ways to calculate the optimal bottleneck values given your measured speeds.

By setting your SQM-enabled router as a slight downstream bottleneck, that means that your router will notice congestion before the ISP's router does, so your router's SQM algorithm can kick in to control the congestion before bloated buffers can build up on the ISP's router. If the ISP's router never sees congestion, it can't bloat up its buffers, so it won't cause the latency (ping time) spike.

I know it hurts to give up even a few percentage points of bandwidth on a slow DSL connection, but fixing the latency problem ends up making your connection feel a lot faster and way less frustrating. People who try this agree almost unanimously that it's worth the trade-off.