eValid™ -- Automated Web Quality Solution Browser-Based, Client-Side, Functional Testing & Validation, Load & Performance Tuning, Page Timing, Website Analysis, and Rich Internet Application Monitoring.

Introduction
This page describes how to organize a LoadTest scenario to ramp up server load in organized steps. In practice, ramping up of server load is used to investigate how the server responds to increased load and to study what happens when the current load is suddenly increased. For example, you may want to start with 10 simulated users, then ramp up to 20 simulated users, and then ramp up to finish with 50 simulated users. Both the steady state response and the response to the rapid increase in users will shed light on overall server capacity.

The easiest way to think about ramping up a scenario is to imagine the fully ramped scenario but then to establish delayed starts to limits the number of users during the beginning phases of the experiment.

Relevant Commands
To delay start of an _eValid command you can use these commands in your eValid *evl script.

• Delay Command.
  The Delay msec command will introduce a fixed delay period that is not affected by the wait time multiplier or wait ceiling settings.

• WaitMod Commands.
  The WaitMod Command Family provides for delays that are based on the current clock time and date for the particular machine from which you are running eValid. You can use these commands to coordinate simultaneous launch of multiple users at a particular clock time.

• User Interaction.
  Another alternative is to use a MessageBox command that gives the user the option of when to proceed with the next ramp.

Example LoadTest Script
Here is andexample of a LoadTest script that (for simplicity) ramps up from 2 to 4 to 6 to 8 simulated users. Each simulated user playback is repeated a total of 50 times; this value should be chosen to make sure early-starting playbacks are still running when late-starting playbacks are launched.

Note: Your actual LoadTest script for a larger number of users would appear the same but would involve many more lines (one for each simulated user). This simplified sample LoadTest script is for illustrative purposes.

  # LoadTest Ramping Example Script
  # eValid V9 
  # (c) Copyright 2012 by Software Research, Inc. 
  # 
  ProjectID "Documentation"
  GroupID "Loading"
  TestID "Rampup_Example"
  LogID "DEFAULT"
  
  Serve FULL
  LoadType FULL
  
  # Start up immediately...Users #1,2...
  
  _eValid "script.evs" "Sample User" 50 "$NAME=value" "-pm 1.0" 1  
  _eValid "script.evs" "Sample User" 50 "$NAME=value" "-pm 1.0" 1  
  
  # Wait 5 minutes before starting the next ramp...Users #3,4...
  
  Delay 300000
  _eValid "script.evs" "1st Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  _eValid "script.evs" "1st Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  
  # Wait until the next quarter hour before starting 
  #    the next ramp...Users #5,6...
  
  WaitModMM 15
  _eValid "script.evs" "2nd Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  _eValid "script.evs" "2nd Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  
  # Wait until the user says OK by clicking on the OK button 
  #    before starting up the next ramp...Users #7,8...
  
  MessageBox "Click OK to proceed to final ramp step."
  _eValid "script.evs" "Last Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  _eValid "script.evs" "Last Ramp" 50 "$NAME=value" "-pm 1.0" 1  
  
  # End of LoadTest Example