Take the 2-minute tour ×
Software Quality Assurance & Testing Stack Exchange is a question and answer site for software quality control experts, automation engineers, and software testers. It's 100% free, no registration required.

Context: Our desktop application communicates with a SQL server, some clients use a WAN connection to reach this SQL server. When they complain something is very slow we ask a list of ping times to analyse the issue. After a developer fixed the issue, I want to verify that it indeed fixes the problem.

With a reported average ping time of 150ms, I use TMNetSim to simulate this on my local machine with these settings: enter image description here This results in similar effects, but still the client is reporting a magnitude of extra seconds than my local test show. In an ideal world I would goto the client location, but since its on the other side of the world this is not really an option :)

Question: What information and setup do I need to get a more realistic simulation?

share|improve this question

1 Answer 1

up vote 1 down vote accepted

This is a very common scenario. Sometimes there are ways to increase the speed of the requests by identifying problems with the route taken, but that requires traceroute and working with network admins and internet providers and is often a bit of a black hole. Often you can not fix the latency issues, so you need to fix your app to function as well as it can without timing out or breaking under those conditions.

One cheap/easy suggestion: find a way to share desktop, or remote desktop to a server at the location that is having issues so you can troubleshoot more.

To specifically answer your question, there is more than just latency to look at and those other things can often play a large role in overall performance in certain geographies, or with poor internet connections (Very common in parts of China and most of Europe):

  1. latency
  2. packet loss/packet corruption
  3. variation/jitter (just because average is 150, doesn't mean there aren't spikes)
  4. bandwidth

There are a number of tools you can use to measure these things, here's an example of one free one: http://www.firewall.cx/networking-topics/general-networking/970-network-performance-testing.html

Once you have that information, then to reproduce those conditions you can plug it in to a WAN simulator such as the one you're using, however keep in mind that they are not all created equal. The most accurate wan simulator I know of is a piece of hardware called a network nightmare: http://www.networknightmare.net/gigEnn/#Products

share|improve this answer
    
Thanks, great answer. I was hoping more for something in the line of don't forget the load on the SQL server, its more than just the network and this is my experience. But latency is of-course mainly network related. Maybe I asked my question wrong :) –  Niels van Reijmersdal Mar 14 at 20:06
    
I'm not sure I fully understand what you're looking for. If you have a more specific question, or could provide more details, I would love to try to answer it for you. If you don't reproduce the issues locally, even when simulating latency and load, then you most likely have 1 of 2 issues. Either you're running different scenarios than they are, or the WAN sim is not simulating close enough to reality. –  Sam Woods Mar 15 at 5:18

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.