Splunk holds its data in a proprietary file "database", structured smartly for efficient searches. In larger deployments, Splunk scales horizontally with many Splunk Indexer instances, each holding a fraction of the data (and potentially replicated copies...).
Here's one way of approaching the docs to understand scaling Splunk for lots of data: http://docs.splunk.com/Documentation/Splunk/6.2.4/Deploy/Distributedoverview
Splunk holds its data in a proprietary file "database", structured smartly for efficient searches. In larger deployments, Splunk scales horizontally with many Splunk Indexer instances, each holding a fraction of the data (and potentially replicated copies...).
Here's one way of approaching the docs to understand scaling Splunk for lots of data: http://docs.splunk.com/Documentation/Splunk/6.2.4/Deploy/Distributedoverview