So I've been using PingPlotter to help troubleshoot connection issues clients are having with our web application. I've had a client run PingPlotter all day and the latency is ridiculously huge. Much higher than I am sure it actually is given speedtests I've run.

Based on feedback from you fine folks here, it seems to me that I'm saturating the clients bandwidth. My question then is: Is the data I've gathered so far worthless, or can I make any assumptions based on the variations between latency during different parts of the day even though that latency numerical value is in fact much higher than it should be.

I've attached my data. Note that normal latency, based on a speedtest going to the same location that PingPlotter was, was 115 ms.

Hey Mike,

Good to hear from you!

*Most* speed tests send out a single ping (or a very few pings) to a speed test server before beginning to download/send out data. That being said, the ping results you get from this are pretty anecdotal - and you can't infer *too* much from them.

There are several instances in the PingPlotter data you provided where the latency matches up to what you're seeing in your speed test tool... as well as instances where there's a pretty large difference. The thing to keep in mind here is that PingPlotter is sending out packets *continuously* in order to give you an idea of the "big picture" as to what's happening with your network connection. If you run the speed test every 15 seconds (the same trace interval that you've got your PingPlotter instance set to here), odds are you'll see a pretty similar variation in the ping results.

If you turn on your "Min" column in PingPlotter (right click in the upper graph, select "Customize View" - and turn on the "Min" option), you can see in your most recent data that, at it's best, the normal latency for this connection is actually around 36ms. That being said - the data that you've got here is far from worthless (and you can infer *quite* a bit from it). Something is definitely saturating the network, but I wouldn't be so certain that you're to blame for this (it *looks* like your issue starts somewhere between hop 2 and 3).

It looks like most of the results that you've provided in previous posts have been during "business hours" (somewhere between 8am-4pm). What happens if you run a PingPlotter trace for a full 24 hours? Does the pattern persist consistently - or do things seem to get better during certain time periods? If you can see a time based pattern in your results, you can use that to start to identify any potential culprits.

Hopefully this helps get you pointed in the right direction! If you have any questions, or need any other help - you know where to find us.

Best wishes,

-Gary