Replication factor

Setting a replication factor is how you specify the number of raw data copies of indexed data you want to maintain across the indexing cluster. Indexers store incoming data in buckets, and the cluster will maintain copies of each bucket distributed across the nodes in the indexing tier (as many copies as you specify for the replication factor) so that if one or more individual indexers go down, the data still resides elsewhere in the cluster. This provides both the ability to search all the data in the presence of one or more missing nodes, and to redistribute copies of the data to other nodes and so maintain the specified number of duplicate copies. 

The indexing cluster can tolerate a failure of (replication factor -1) indexers (or peer nodes, in Splunk nomenclature). If you are using a replication factor (RF) of two, the cluster maintains two copies of the data, so you can lose one peer node and not lose the data altogether; if you use an RF of three, you can lose up to two nodes and still maintain at least one copy; and so on.

The trade-off is that your cluster will need to store and process multiple copies of data. The replicating activity doesn't consume much processing power, but as the replication factor is increased, you will need more indexers and more disk storage for the indexed data.