Link to home
Start Free TrialLog in
Avatar of Axis52401
Axis52401Flag for United States of America

asked on

ISP latercy troubleshooting

This may be a vague question but we have an office with 2 ISP connections Cable and DSL. We're testing a remote hosted app that is hosted by a third party somewhere else. It connects but users report it stalling and disconnecting. I have tried isolating the users of this app to one or the other ISP with firewall rules to no effect so it doesn't seem to matter which ISP we use. The vendor put a monitoring program on one of our PC's and basically all ti does is pings google.com and its app url and track the statistics. Its showing almost 20% packet drop in their log. When I set a continuous ping to google or their IP and let it run I see an occasional no reply maybe 1 out of 100 but not 20% so I don't know how its getting that. Another thing the ping times to their IP address vary from about 50ms to 3-500. Their support says anything over 200ms will cause these symptoms.  So my question is how can I trouble shoot this problem. Both ISP's report everything is working fine.
Avatar of Mandeep Khalsa
Mandeep Khalsa

I was in a similar situation recently with one of my clients. They had 2 ISP's and both were plugged into a single switch which then was routed to multiple firewalls and the rest of their network. This switch in the front was used because they wanted to do failover and this was the best way to plug them in. However this switch had crappy backbone and it would only work on a certain speed/bandwidth/latency etc. It was replaced with a newer switch and things improved a lot.

Since your question doesn't tell us what your network environment is like I can't be specific on what you should try however if you can get something out of my recent experience it would be a win win situation.
I suspect that your dual internet connections are configured for load balancing, See if you can configure them in fail over mode instead. Your issue is probably that the paths (number of hops thus latency) to the third party host is significant enough that if your network establishes a connection on ISP A then routes subsequent request via ISP B that is when you hit the issue.  Quick test would be to disconnect your slower ISP and run only via the fastest this would become your primary connection in the failover configuration.
Avatar of Axis52401

ASKER

The ISPs go into the firewall. They are configured for fail over not load balancing. I have tried routing the traffic to this application through both ISP connections indivudually.
Have you compared the ping results from the computers as well as from the firewall itself? If they are different then your switches internally would be the culprit here.
I've tried it from various computers but we have a sonicwall and I don;t see a way of doing a ping from the firewall itself
ASKER CERTIFIED SOLUTION
Avatar of Mandeep Khalsa
Mandeep Khalsa

Link to home
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
Start Free Trial
OK, found it, it pings 66ms but I can't set a continuous ping to try it more then one at a time