Measuring Web Content Accelerator Changes

This section explains how to test Web Content Accelerator and measure Web page performance using a variety of tools. These measurements can be used to determine how effectively Web Content Accelerator is accelerating your Web service.

Checking That Web Content Accelerator Is Active

To check that your Web content is being accelerated, use your browser’s “View HTML Source” option after loading a Web page. Search for the following comment:

<!-- Optimized for speed – Web Accelerator -->

If you do not see this comment, see Configuring Web Content Accelerator for Your Services.

Using Stealth Mode to Test Web Content Accelerator

If you do not have a development or test environment available you can still test Web Content Accelerator by running the accelerator in Stealth Mode.

Stealth Mode is not available with Web Content Accelerator Express.

In Stealth Mode, Web Content Accelerator will not apply any optimization to Web pages, unless commands are passed within the query string. End users will not see any change in the way Web pages behave or appear.

When running Web Content Accelerator in Stealth Mode all incoming requests for pages and objects will be served in their original form. If the request contains the parameter aptimizer=on in its query string, Web Content Accelerator will send a cookie that represents an explicit instruction to accelerate all requests for this browser session.

When Web Content Accelerator sees the enabling cookie on the incoming requests it will intercept and accelerate them according to the rules and settings defined in the Web Content Accelerator profile applicable to that scope.

These are examples of valid requests:

www.example.com?aptimizer=on
www.example.com/destinations?aptimizer=on

You configure Web Content Accelerator to run in Stealth Mode by changing the aptimizer_mode configuration key in the relevant profile.

Measuring Web Page Speed

To find exactly what has changed since enabling Web Content Accelerator you will need before and after Web page load time measurements.

An initial measurement should be used as a baseline to which further measurements will be compared. Ideally the baseline should be documented while the Web site is not yet running with Web Content Accelerator enabled, or with Web Content Accelerator running in Stealth Mode.

At each stage a new set of measurements can be taken and used for comparison with the baseline. The stages could be any milestone in the Web Content Accelerator deployment, for instance:

Initial deployment with default settings

Changes to rule options

Changes to CSS settings

Changes to caching behavior

Changes to default image sets

Tools

There are many tools available to verify Web page load times and provide feedback on Web page structure. Some can be used in a very granular fashion, allowing testing on individual Web pages. Others are used for collecting metrics that can be correlated to the Web Content Accelerator-enabled implementation.

Page Speed and Structure

Many tools can test and provide information about your Web pages. Some are available as online services while some are available as an add-in for Web browsers such as Firefox or Chrome. Some typical examples follow, however your exact requirements may vary:

WEBPAGETEST (www.webpagetest.org) is a free web site service that emulates browser behavior and retrieves your web page for analysis, presenting results in a cascade-style chart showing how a browser progresses through the page when rendering it.

A simple process for ensuring that the site is as fast as possible is to work your way down the stepped list of objects on the chart, determining the most appropriate acceleration to apply for each object. For example, the start of the chart might show two redirects and then the page loading. The ideal action for this would be to rework the site such that the redirect is not used. Or you may be looking further down the chart at a script that loads on the page.

YSlow (http://developer.yahoo.com/yslow/help) is a browser add-in developed by the Yahoo! Developer Network initiative. The software uses built-in rules to grade a web page across a range of performance indicators, from A (best) through F (worst) and provide suggestions to improve the web page score. It runs on Firefox and Chrome browsers.

Google Page Speed (http://code.google.com/speed/page-speed) is another browser add-in also available for Firefox and Chrome browsers. The software grades web pages on a Page Speed Score from 0 (worst) to 100 (best). Google Page Speed also provides suggestions on how to improve the web page score.

Other Metrics

Some metrics require a longer data collection period to be useful. This means collection must start even before Web Content Accelerator is enabled.

Google Analytics (www.google.com/analytics) is a free tool that can be used to collect metrics such as page load time, number of page views, page views per visits and bounce rate. Google Analytics requires a small script to be added to all pages in your web site. Results are visible a couple of days after installing the script, but you need to collect data for some time to allow for effective comparisons.

Google Webmasters (www.google.com/webmasters) is another free tool that can be used to manage how Google sees your web site. As a bonus it provides a chart showing the average global web page load time for your entire web site, over a rolling window of six months. Google Webmasters requires site ownership authentication, which can be accomplished by a small DNS change, a configuration file added to your web server, or a meta tag added to your default web page.