WHAT'S IN THIS CHAPTER?
Looking at how ESP scales up across multiple machines to handle more content and searches
Understanding ESP subsystems and how they scale
Understanding the effects of networked storage
Managing virtualization and ESP
Working with multinode installations
Working with Linux and Solaris installations
Using the Document Processing Pipeline
This chapter defines the basic components of Microsoft FAST Enterprise Search Platform 5.x. FAST ESP. The main focus is understanding multinode installations. Why and how are extra machines added to an ESP installation, and how are processes allocated? Multinode setup is the cornerstone of most ESP production installations. Additional configuration for Unix and Linux systems are considered.
FAST ESP is a highly scalable, robust search platform that can be customized. In addition to scalability, ESP can be configured to handle virtually any type of data, given enough customization effort. On the front end ESP offers virtually every results list navigator in the industry, although those can also take some effort to fully utilize, and very few search applications would ever use all of them.
Virtually all production FAST systems are multinode, for both scalability and reliability. Even for a small number of documents and light search traffic, production search systems are usually deployed on a minimum of three nodes to allow for failover.
In the ...