I am confused which approach would be better having single cluster with 12 nodes or having 3 cluster with 4 nodes each in elastic stack. What are the advantages and disadvantages of single cluster? Does elastic charge me for 3 cluster as far as I know they charge for nodes but can someone clarify which would be better approach and which would be cost effective solution?
Since you want to use ML (based on your node type list) - then you will require at least a platinum license.
In general, you can choose whether or not you want to buy node licenses or "capacity-based" licenses. To get "capacity-based" licenses, you need to license at the Enterprise tier, where you buy blocks of ERUs (explained below).
What's the difference?
In the node model, a licensed node is an instance of the following node types, each of which is up to 64GB in size:
** Data nodes
** Dedicated Master nodes
** Nodes used for cross cluster operations
** Dedicated Machine Learning nodes
Keep in mind that a 1 GB data node is taking up the same license as a 64GB data node. The number of clusters in the node-based model sort of does matter. We'll want to have a separate license for each cluster that's in production (non-prod clusters are not licensed).
In the "capacity-based" model, an ERU (Enterprise Resource Unit) is defined as block of 64GB of capacity. You are free to create one 64GB node, eight 8GB nodes, sixty-four 1GB nodes, whatever. The number of clusters created also doesn't matter.
In general, despite the fact that an ERU at the Enterprise Tier is more expensive than a node at the Platinum tier, the concept of an ERU allows you much more flexibility and also allows you to use one of our orchestration products to help you manage/deploy clusters (either ECE or ECK)
Apache, Apache Lucene, Apache Hadoop, Hadoop, HDFS and the yellow elephant
logo are trademarks of the
Apache Software Foundation
in the United States and/or other countries.