5 reasons why we love Elasticsearch (And You Should, Too!)
Reason 5: Multitenancy
At the start, Elasticsearch is often used for just one single purpose (i.e. full-test search) and for one department only. Throughout the time we see that other departments and users also see the benefit of using Elasticsearch. To prevent the use of one index for multiple users, Elasticsearch supports multitenancy. This means that we can create an index for each user in the same cluster. So, users will be able to search through their own indices or across multiple indices in a cluster. As from version 5.3 and above Elasticsearch also supports cross cluster search, which gives you the possibility to even search through remote clusters.
Reason 4: Scalability
Because we all have a fast growing amount of data and a constant challenge of having our applications perform well, it is sometimes necessary to add more resources. The easiest way to do this is to scale horizontally (scale-out). Just create a new instance of Elasticsearch on a new server, give it the same cluster and a unique node name, and Elasticsearch will do the rest. This means that it automatically manages the distribution of indices and queries across the same cluster.
Reason 3: It is fast!
One of the mean reasons for companies to choose for Elasticsearch is performance. Elasticsearch is known for searching millions of documents in a few milliseconds. To be able to do this, Elasticsearch uses Lucene under the hood, which is a search and indexing library. Elasticsearch uses so called “word-level inverted indices”, this means that an index contains a reference list for each word and its position in a document.
Reason 2: Ease-of-use
Elasticsearch can be used in just a few steps. Just download the software, unpack and configure it and you are ready to go.
Reason 1: Costs
Because Elasticsearch is an open source solution and free for use. By using a single cluster and one node or multiple nodes you are able to start indexing your data and logs without spending anything, just your time will be sufficient.