Talk about the idea pipeline, and reference implementations.

Talk about various connectors that we have created.

Distributed, fault tolerant, highly available.

Document Orientated

Stages

Pipelines

The Easy Connectors – CSV, XML, JSON, text, rss, etc.

The Hard Connectors – enterprise systems such as documentum, salesforce, and sharepoint.

Incremental РAlways attempt to detect when items change and only send the changes.  Avoid leaving orphaned search results!

Security and Permissions – While indexing, the need to obey the underlying security model is paramount for the enterprise.