I received several ping plotter results from a client who has been experiencing slowness in a hosted application. After having a colleague re-evaluate the results, it seems pretty evident there are latency issues in the client's local network. The hosted application relies on Citrix to conenct to the datacenter, so anything over 100 ms ping times can be considered a performance issue, especially when there are several active users.

The Ping plotter was too big to attach, so I've taken a couple of Jing Screenshots below. What you're seeing there is a consistent pattern:

1. Starting at 10:24 AM: http://screencast.com/t/bmoYZ9JRSWvx
2. Starting at 10:09 AM: http://screencast.com/t/d6rJ30b22t
3. Starting at 9:42 AM: http://screencast.com/t/7FJ97kxKk1Kv
4. Starting at 9:06 AM: http://screencast.com/t/fuZCgMiQKyVj
5. Starting at 8:27 AM: http://screencast.com/t/R4zpZlf1aUxA
6. Here's an example of extreme latency starting at early in the morning (5:46 AM), for example: http://screencast.com/t/HCeah2Ejw6hN
a. Here's a zoomed in version of that. We can see a direct correlation between the first hop and the final destination: http://screencast.com/t/e7XMUHMn. I believe this indicates that the first hop could be the root cause of the slowness.



Now, the client claims to give low priority to ICMP traffic. I can understand if that was evident in the ping plotter by Packet loss, but even if it were given priority, wouldn't the latency still show up?

Thanks!

Hey Bobby,

Thanks for writing in!

You've done some excellent detective work here! There are definitely some patterns (where the latency spikes) occurring at hop #1 that are carrying through to the final target. This would make for a strong case that the culprit in these issues could be somewhere between the user and hop #1.

What is the device that's at hop #1 - and how is this machine connecting to the network (wired or wireless)? Usually anything inside of a user's network is going to return a response of 1ms or less, but it looks like the average here (per the .pp2 files you emailed over to us) is closer to 6ms (which *may* be a bit of a red flag, depending on their setup...).

The goal at this point should be to isolate and eliminate any variables (that you/your client have control over) to see if you can definitively narrow down what the source of the problem is here here. If you (or your client) are interested, we've got a guide that goes over a few different strategies on this front, which you can find here:

http://www.pingplotter.com/netnirvana/

If your client is claiming that they're giving low priority to ICMP traffic, you could always switch over to a different packet type (TCP) to see if that has any impact on the results as well.

Hopefully this helps out! If you should find yourself with any questions, or needing ay further guidance - don't hesitate to let us know.

Best wishes,

-Gary