Demonstration of an Applet Viewer for Performance Load Testing/Modeling Results

This brief tutorial demonstrates features of a Java Applet Viewer, which facilitates user interaction with performance load testing/modeling results and charting options. Using the Viewer controls, the user can scan through different values for input parameters and instantly see how performance curves are affected.

Example 1. WebLogic 8.1/Grinder Load Test - Throughput per Second (TPS) vs. Workload Threads

This example is taken from a load testing study with additional details regarding the relationship of Execute Queue ThreadCount and max-beans-in-free-pool and max-beans-in-cache configuration parameters of a WebLogic 8.1 Server.

In this example, only one of the five statistics in the Statistics panel is selected for viewing. Its label, Grinder TPS, has the same color as is use to draw its performance curve. Its units of measure, transactions/second, are labeled on the Y axis.

In the Input Parameter panel there are eight choices. Clicking on the horizontal split pane separator and dragging it upwards you will see that Grinder_Threads is selected as the primary independent parameter (indicated by the X1: prefix). The label for this parameter also appears on the X axis of the chart along with its units of measure, Agent Threads.

The user can interact with the checkboxes, label selections and split pane controls to change the presentation of the chart and its curves. The following are step-by-step instructions for interacting with the Viewer.

Full Chart View The vertical split pane separator between the controls and labels on the left and the chart on the right has left and right arrows at the top that can be used to collapse and expand the Statistics and Input Parameter Panels. Click the left arrow at the top of the separator to cause the chart to occupy the full width of the applet Viewer and add a statistics legend list in full chart view.

Click the separator's right arrow to make the control/label panels visible once again. Also, the vertical split pane separator can be dragged to change the width of the chart.
Full Statistics/Input Parameter Panel Views Similarly, the horizontal split pane separator between the Statistics and Input Parameter panels can be used to collapse, expand or change the sizes of these panels. Scroll bars bring specific sections of the labels in these panels into view.
Selecting Statistics to Chart Clicking on the label(s) in the Statistics panel performs several functions. If the label was not previously selected, clicking on it will cause the label to be color-highlighted and its statistic added to the chart. Clicking on a selected label removes its color-highlighting and its statistic from the chart. Toggling a statistic's label from ON to OFF to ON will select a different color for the label and its curve in the chart. In other example, we'll see how multiply selected labels are grouped together at the top to the panel.
Selecting Primary Input Parameter Selecting the X1 Radio Button allows the primary input parameter to be selected. In this mode, clicking on a label in the Input Parameter panel will change the input parameter used for the X axis. Notice that the color-highlighting changes for the label as does the X Axis labeling and scaling.
Selecting Secondary Input Parameter Selecting the X2 Radio Button allows a secondary input parameter to be selected. In this mode, clicking on a label (other than the primary label) in the Input Parameter panel will select a secondary parameter if one was not previously selected, or change from one secondary input parameter to another if one was previously selected. Or, de-select a secondary parameter if a previously selected label is clicked again. Click on an input parameter label and notice that the color-highlighting changes for the label. Also, notice that the X Axis labeling now has the name of the X2 parameter added after that of the X1 parameter. Below and associated with the X2: parameter's name is a legend box with the range of values, which the secondary parameter can have. Each value has a symbol that replaces the square as the datapoint indicator for a curve having that value. If there was one curve before, there will now be as many curves as there are X2 values to chart.
Animate Checking the Animate checkbox with a Secondary Input Parameter selected causes the Viewer to chart one curve at a time from a family of curves with different X2 values. Only the symbol in the X2 legend box for that curve is show for the X2 value associated with that curve. Autoscaling is turned off in the animate mode.
AutoScale With the AutoScale checkbox checked, the X and Y scalings are adjusted so that all of the curves' data are visible in the chart. Toggling AutoScaling OFF does not have an immediate affect on the chart's appearance whereas toggling it ON can trigger an immediate re-scaling of the chart, depending on intervening changes to other controls and labels. A primary benefit of turning AutoScaling OFF is realized when using the X2+/- control to increase or decrease non-primary input parameters' values. With AutoScale OFF, users can more easily see statistics curves change with changes to those input parameters.
Changing Input Parameter Values Selecting the X2 +/- Radio Button allows any non-primary/non-secondary input parameters' values to be examined and changed. In this mode, some of the possible values for a parameter (other than the primary and secondary parameters) are shown on its label as the mouse cursor is moved over the label without clicking on it. The possible values shown on the label are its current value surrounded by the < > symbols and the next higher and lower values if they existing. Clicking on a label causes its parameter's value to be increased to the next possible value if one exists. Right mouse button clicking or Shift-Left mouse button clicking causes the parameter's value to be decreased to the next possible lower value if one exists.

Next: Charting Multiple Statistics

Copyright © 2014 Anser Enterprise

Uses JFreeChart Open Source Charting Library