FAST ESP Configurations - partner briefing

Deployment configurations:

To reduce complexity Crawler, Connector, Anchor, Content Distributor and Indexing Dispatcher nodes are omitted.

3 server configuration:

Let us assume that the customer only has budget for 3 servers. Having a 3 server configuration the Search cluster is arranged as a single column comprising 1 indexing row and 2 search rows. This configuration:

- Provides limited capacity, but distributes to the best of its ability the load.

- Search Redundancy is provided by duplication of Search nodes and Query/Result nodes.

5 server configuration:

The addition of 2 extra servers offers some more overall capacity compared to the previous configuration. We can now duplicate the Document Processor and Indexer nodes for redundancy together with Query & Result redundancy.

Depending on the arrangement of the SEARCH CLUSTER the configuration provides either redundancy for these components, or more indexing capacity enabling a larger volume of documents to be handled.

In the Search & Index Redundancy model. SEARCH CLUSTER = 1 COL 2 ROWS . Here Indexers are Active-Passive. We have one master and one backup indexer.

In the Increased indexing capacity model. - > SEARCH CLUSTER = 2 COLS 1 ROW - Indexers are Active - Active.

Advantage: Backup takes over immediately on failure.

Disadvantage: Indexing on all rows is performance heavy. This reduces search performance as we have less dedicated search servers. Adding additional search servers increases hardware costs.

7 server configuration:

With 2 more Servers, we have a 7 Server configuration:

Now we can split the Document Processor and Indexer to free up resources for each of them.

Redundancy is provided for all components involved in document processing, indexing and search.

This will give us:

2 hosts running Document Processor nodes.

2 hosts running Indexer nodes.

2 hosts running Search and Query & Result nodes.

1 host running Administration (Admin) node.