Professional Apache Tomcat 6

Chapter 17. Clustering

The Tomcat server has grown up, from a reference implementation of a Servlet container for demonstrating and testing new APIs to a robust and high-performance Web-tier server. Increasingly, Tomcat is being used in production scenarios to handle real-world Web applications. It is a prime example of a prototype becoming the product. In a sense, the designers and architects realized that their mission changed direction, and they redesigned the Tomcat product for high-stress production deployments.

Tomcat 6 is further along the evolutionary path of improved performance. Real-world deployments place many tough demands on the Tomcat server. Many of these requirements were not important to the Tomcat development team when it was a mere reference implementation. Two such areas are support for horizontal scalability (the ability to handle increasing user requests by utilizing a group of physical machines) and high availability (the ability to survive hardware or software failures and maintain a high percentage of application uptime). In other words, what happens when there are so many users that Tomcat servers start to crash, and how can you ensure that no user loses data? These are real-world problems, and Tomcat 6 attempts to solve them by providing built-in support for clustering. Clustering, in this context, refers to running multiple instances of the Tomcat server so that it appears to users as a single server.

This chapter covers the many facets of Tomcat 6 clustering, ...