Chapter 9. Deploying Storm on Hadoop for Advertising Analysis
In the previous two chapters, we saw how we might integrate Storm with a real-time analytics system. We then extended that implementation, supporting the real-time system with batch processing. In this chapter, we will explore the reverse.
We will examine a batch processing system that computes the effectiveness of an advertising campaign. We will take the system that was built on Hadoop and convert it into a real-time processing system.
To do this, we will leverage the Storm-YARN project out of Yahoo! The Storm-YARN project allows users to leverage YARN to deploy and run Storm clusters. The running of Storm on Hadoop allows enterprises to consolidate operations and utilize the same infrastructure for both real time and batch processing.
This chapter covers the following topics:
An introduction to Pig
YARN (resource management with Hadoop v2)
Deploying Storm using Storm-YARN