The function of a poller endpoint is to read data, in the form of files, from a location on a file system and pass that information to other endpoints in the ESB. Poller endpoints create an in-only message exchange containing the data read in from a file.
A poller endpoint, as its name implies, works by continually polling the file system to see if a file is present for consumption. The polling interval is completely customizable.
You can also control the files a poller endpoint consumes. Using the basic configuration attributes, you can configure the endpoint to poll for a specific file or you can poll it to monitor a specific directory on the file system. In addition, you can configure the endpoint to use a custom file filter.
By default, poller endpoints will only consume valid XML files. You can customize this behavior by configuring the endpoint to use a custom marshaler.
Poller endpoints play the role of consumer from the vantage point of the other endpoints in the ESB. As shown in Figure 2.1, a poller endpoint watches the file system for files to consume. When the endpoint consumes a file, it transfers its contents into a message and starts off an in-only message exchange. Poller endpoints cannot receive messages from the NMR.
Poller endpoints are configured using the poller element. All its configuration can be specified using attributes of this element.
The more complex features, such as custom marshalers, require the addition of other elements. These can either be separate bean elements or child elements of the poller element.
 |  |  |
 |
