Chapter 44. Implementing a Processor


Abstract

Apache Camel allows you to implement a custom processor. You can then insert the custom processor into a route to perform operations on exchange objects as they pass through the route.

44.1. Processing Model

Pipelining model

The pipelining model describes the way in which processors are arranged in Section 5.4, “Pipes and Filters”. Pipelining is the most common way to process a sequence of endpoints (a producer endpoint is just a special type of processor). When the processors are arranged in this way, the exchange's In and Out messages are processed as shown in Figure 44.1, “Pipelining Model”.

Figure 44.1. Pipelining Model

Pipelining model
The processors in the pipeline look like services, where the In message is analogous to a request, and the Out message is analogous to a reply. In fact, in a realistic pipeline, the nodes in the pipeline are often implemented by Web service endpoints, such as the CXF component.
For example, Example 44.1, “Java DSL Pipeline” shows a Java DSL pipeline constructed from a sequence of two processors, ProcessorA, ProcessorB, and a producer endpoint, TargetURI.

Example 44.1. Java DSL Pipeline

from(SourceURI).pipeline(ProcessorA, ProcessorB, TargetURI);
Red Hat logoGithubRedditYoutubeTwitter

Learn

Try, buy, & sell

Communities

About Red Hat Documentation

We help Red Hat users innovate and achieve their goals with our products and services with content they can trust.

Making open source more inclusive

Red Hat is committed to replacing problematic language in our code, documentation, and web properties. For more details, see the Red Hat Blog.

About Red Hat

We deliver hardened solutions that make it easier for enterprises to work across platforms and environments, from the core datacenter to the network edge.

© 2024 Red Hat, Inc.