Cascading is an application development and a data processing framework that makes Hadoop development simpler and more durable over time. It does this by reducing the number of lines of code that are required to perform most tasks, and durability comes through its seamless support of Hadoop and other big data frameworks.

Hadoop application development with MapReduce and other frameworks is a difficult task with a very steep learning curve, as you have seen in the previous chapter. Cascading significantly simplifies this process with an intuitive high-level abstraction over Hadoop application development, job creation, and job processing. With Cascading, developers can create reusable components and pipe assemblies, which are reliable and scalable.