LinkedIn is a high traffic consumer internet site and LinkedIn search serves millions of queries per day.  The search problem at LinkedIn is unique in several ways:

1) High volume with both unstructured and rich structured data

2) Different types of structured data, e.g., the social graph

3) Faceted search on the structured data

4) Real-time changes to data

5) Constant expansion of the underlying corpus

This session will discuss LinkedIn’s distributed real-time faceted search system, which is designed to guarantee strict SLAs for both search latency as well as how rapidly a newly updated document appears in the search results.  It will further discuss incremental index updates, query segmentation and classification, as well as various caching approaches.