Elastic Search

Elastic Search 

Highly scalable open source search and analytic engine.Store search and analyse big volume of data.

Written in Java. 

• NTR - Near Real time - From the time you index a document to the time its searchable. With a delay of 1s
• Cluster - collection of nodes(servers) that hold the data. Cluster is identified by a name and nodes join the cluster by its name. 
• Default name - elasticsearch
• Node - Single server part of cluster stores data 
• Default name - UUID
• Index - collection of documents with similar characteristics..  - Database. 
• there can be multiple indexes in a single cluster.
• Type - Category of multiple documents. index has one ore more types.  Logical partition of your index. Name and mapping. Similar to mysql table ? 
• Class in OOP.
• Index can have one or more types.
• Document - a unit of info. Json Schema less . Stored with in an index / type
• Key - value pairs.
• Object in OOP.
• Mapping - Similar Schema of table for a document. 
• Include data type 
• How fields should be indexed.

Analogy to Relation Database :

1. Index - DB
2. Type - table
3. document - row in table
4. Fields - column 
5. Mapping - schema 

Shards  - Improve scalability 

Index divided into multiple pieces called shard.

Improve scalability.

Index contains more data than the node can store. Sharding comes handy.
Allows distribute and parallelise operations.

Replica : 

Copy of Shard.
Handles node failure.