appserver.io introduces the concept of so-called webapps as a new naming for PHP applications. This new naming is used to differentiate between what we call legacy applications, PHP applications as of the current generation, and applications specifically developed for the appserver.io ecosystem.

A webapp is a PHP application that can only run within the appserver.io infrastructure as it heavily relies on services provided by it.

The definition above means that the application uses servlets and must therefore be executed within a Servlet Engine .


A webapp needs a certain folder structure. We will describe this structure and its background.

Any webapp may contain three directories that are expected to have a certain functionality. These three are the following.

Name Description
META-INF Files with relevance to background services such as the Persistence Container .
WEB-INF Files handling incoming requests and a relevance to services responsible for user interaction such as the Servlet Engine .
common Files which might be needed by both.

The reason for this separation will be given further below.

All three directories expect a certain structure for their contained files:

  • Classes and other PHP structures MUST be located within a classes folder and follow a PSR-0 compatible folder structure
  • All appserver.io specific configuration files MUST be placed as direct children of the structure folder
  • All other files can be structured as needed

This results in a folder structure such as the following.

    |- common
            |- classes
                    |- <PSR-0 path to classes>
    |- META-INF
            |- classes
            |       |- <PSR-0 path to classes>
            |- <Configuration files>
    |- WEB-INF
            |- classes
            |       |- <PSR-0 path to classes>
            |- <Configuration files>
    |- vendor

The structure above holds true when using the default configuration of the appserver. If this structure has to be changed for some reason, it can be done by using the application configuration

Responsibility separation

As mentioned above, the webapp is separated into different directories. The directory separation into META-INF and WEB-INF (and common as common ground) serves the purpose of responsibility separation.

To put it simply, we can state that files from META-INF get processed by backend services, and WEB-INF by services responsible for client interaction.

The purpose of this separation is hinted at by the use of the term "service" and prepares for a situation in which services are spread over a network.

When having a closer look at app construction one can see that servlets (WEB-INF) and processors (META-INF) communicate over a proxy object. This proxy allows to communicate using remote method invocations and supports scalability by allowing for asymmetrical scaling of application services.

Thus, a central Persistence Container could handle several server instances serving servlets and static files to connected clients.

Load Comments

Please note that by clicking "Load Comments", your browser will establish a connection with servers from Disqus.
For more information, please visit our privacy statement.