AEM 6.5 Dispatcher Experiments

This repository contains a collection of Dispatcher experiments in take-home lab format. The content is intended for intermediate to advanced AEM developers and customizers.

Goal

The goal of these experiments is to raise awareness about the Dispatcher, and provide a project for test driving its feature set.

Non-Goals

This repo does not attempt to prescribe a one-size-fits-all configuration for the dispatcher. Due to the myriad use cases that AEM supports, it would be impossible to do so. Instead, pick and choose concepts from the below experiments and try them out on your project.

Getting set up

You will need the following SDKs, tools, and apps installed to work through the experiments:

• Java 11.0.*
• Maven 3.6.*
• Node 10.15.*
• JMeter
• VisualVM

You will also need a local AEM author/publish/dispatcher setup:

• AEM 6.5 author instance running on :4502
• AEM 6.5 publish instance running on :4503
  • Ideally, both instances will have the latest Service Pack installed
• Dispatcher accessible at aem-publish.local:8080
  • macOS Catalina & Mojave users can find macOS dispatcher setup instructions here
  • Windows users can find IIS/Windows 10 instructions here
• Cache flushing agent configured on the Publish instance

You will need to install the AEM project code contained in aem-project/ on your local author and publish instances:

• Ensure your Maven settings.xml file includes a profile for the Adobe Maven Repository
• Build and install the project code on your local author and publish:

cd aem-project/
mvn -PautoInstallSinglePackage -PautoInstallSinglePackagePublish clean install

Ready? Confirm by checking:

• Navigate to http://aem-publish.local:8080/content/dispatchertester/us/en/regular-page.html. The rendered page should look like this:

• And, the dispatcher docroot (cache directory) should contain a static copy of this page:

Experiments

1. Effect of Re-fetching Dispatcher Flush

In the HelpX article Optimizing the Dispatcher cache, the concept of "Re-fetching Dispatcher Flush" is introduced. This experiment shows the impact.

⇨ Effect of Re-fetching Dispatcher Flush

2. Effect of a /statfileslevel greater than 0

The configuration included in the dispatcher download contains a /statfileslevel set to 0. This experiment demonstrates the performance impact that can be achieved by increasing this value.

⇨ Effect of a /statfileslevel greater than 0

3. Effect of an /ignoreUrlParams allow list

Configuring /ignoreUrlParams in an allow list manner is important to improve your cache hit ratio, which in turn can save your publish tier from unnecessary content rendering.

⇨ Effect of an /ignoreUrlParams allow list

4. Effect of the /gracePeriod setting

gracePeriod is a relatively new feature which can reduce spikes in load when a number of cache invalidation events occur in quick succession.

⇨ Effect of the /gracePeriod setting

5. Effect of /enableTTL

Dispatcher versions 4.1.11 and greater can be configured to respect a time-to-live (TTL) based content "timeout". Use this experiment to see how it works.

⇨ Effect of /enableTTL

6. Use of the Dispatcher Optimizer Tool

A tool has been created to:

• read (parse) a Dispatcher and Apache Httpd configuration
• output detailed parsing errors
• create a report indicating areas where the best practices were not followed

The tool can be run on any Dispatcher configuration, and optionally an associated Apache HTTPD configuration.

⇨ Running the Optimizer Tool

Contributing

Contributions are welcomed! Read the Contributing Guide for more information.

Licensing

This project is licensed under the Apache V2 License. See LICENSE for more information.