At times I have the feeling that broadband performance is poor.
Using firefox under windows10, perhaps Gmail will complain it isn't connecting properly, or comments will not load on a website.

Sometimes it really is an indication that the linespeed has dropped unacceptably to half-bb or dial-up speed, sometimes the speed/bandwidth seems reasonable, should not be causing problems.

How can I tell how subjective poor broadband performance issues are due to the modem/router hardware (or possibly cable, connectors, laptop NIC), versus how much is due to intermittents on the line to exchange (and beyond), or possible over-allocation to capacity?

Out of curiosity I used pingman (pro, trial) to monitor the metre or so of cable to my router (router IP).

Surprisingly this is not rock-solid; there is a good average of 1.2mS, but some peaks of up to 3mS, and even some fast pings of 0.2mS. (see screenshot) Is this normal or indicative of breakdown?

Win10 command line >pathping to the router, however, says 0,0,0 mS (microsoft; it may well be lying, but not by much).

I am not seeing actual packet loss (except one router socket has died). I get comparable results when I reverse the cable to use the other set of wires in the cable(cat6), and when I replug at the router.

Monitoring remote sites (G~ DNS at 8.8.8.8, and my ISP), again I get some rare spikes, with a little accompanying jitter. These maybe double or treble the latency. Nevertheless, the signal gets through, without packetloss.

Pathping results look normal.


As I was writing, my connection to external sites was solid, but then connection was lost to one site (brief red bar), latencies grew in tandem, then the situation recovered. Some big jitter at the time. (See screenshots.)


I use speedof.me, a non-JS HTML5 speedtest. My laptop is a fairly old Dell, it has been scanned for malware using ESET online, which should catch most problems. Drivers are latest, so far as I can tell. But it's an old laptop. Current o/s is win10 (generally I would use a linux such as manjaro).

Hey methuzla,

Thanks for reaching out!

The first screenshot you've provided (erratic results to local router) shows some interesting results. If you're pinging the router while plugged directly into it, you'd usually expect to see <1ms average latency (especially if the cable you're using is only about a meter long). If you've got another cable handy - you might want to try swapping this out to see if it has any effect on your results.

As far as the difference you're seeing in the results in PingPlotter and your command line ping - the first thing I'd recommend looking into is what packet type you're using in PingPlotter ("Edit" -> "Options" -> "Engine"). For best results, you'll want to make sure this is set to "ICMP Using Windows DLL (default)." We've got a knowledge base article that goes over some other potential reasons for differences between PingPlotter and the command prompt results (the article uses "ping -t" as an example, but the ideas here can apply to other ping/traceroute methods as well):

http://www.pingman.com/kb/37

I can see the latency spike/packet loss pretty clearly in the other two screenshots you've got; but it's tough to tell exactly *where* this issue might be originating from without being able to see graphs for the other intermediate hops. If you'd like to send us a .pp2 file (or use the "File" -> "Share" option in PIngPlotter and shoot over the link) - we'd be happy to take a closer look at your results and offer any guidance we can from there.

The best way to narrow down your issue is to keep monitoring (continuously) over time to a target that you've been experiencing problems with. Once you've got a good amount of data in PingPlotter, you can start trying to identify any patterns that may be present in your results. You can use that information to start narrowing down potential culprits, and then start isolating and eliminating any variables (that you have control over) to narrow things down from there (all while continuing to trace with PingPlotter - so you can see what does/doesn't have any effect in your results). We've got a great guide that goes over this process (as well as provides quite a few examples) in more detail, which may prove helpful to your cause:

https://www.pingman.com/network-nirvana/

Hopefully this helps out. If you have any questions, or if you should find yourself needing any other guidance - just let us know!

Best wishes,

-Gary