In order to perform searches and other operations on the big data produced by some systems in my daily usage I had to choose between three different and very popular Open Source technologies:
Lets' have a quick look at the features of each one of them first.
Elastic Search
Elastic Search (http://www.elasticsearch.org/) is a search engine based on Lucene (http://lucene.apache.org/). Even if some data visualization and analytics features have been added to it, it is basically a pure search full text search engine into schema free JSON documents.
MongoDB
MongoDB (http://www.mongodb.org/) is a cross-platform NoSQL document-oriented database. Data are stored in JSON-like documents into collection in dynamic schemas. The data interchange format of the documents is BSON (Binary JSON). BSON types are a superset of the JSON types. Like Elastic Search, MongoDB allows full text search in documents, but it provides more useful features like server-side scripts in JavaScript, MapReduce (a data processing paradigm for condensing large volumes of data into useful aggregated results) support, capped collections (the possibility to define a maximum size for a collection: so once it fills its allocated space, it makes room for new documents by overwriting the oldest documents in the collection; this feature could be useful capturing logs or any streaming data for analysis), and an aggregation pipeline framework.
Hadoop
Apache Hadoop (http://hadoop.apache.org/) is a framework for distributed storage and distributed processing of Big Data on clusters of commodity hardware. Its Hadoop Distributed File System (HDFS) splits files into large blocks (default is 64MB or 128MB) and distributes the blocks amongst the nodes in the cluster. For processing the data, the Hadoop Map/Reduce ships code (specifically Jar files) to the nodes that have the required data, and the nodes then process the data in parallel. At present time it provides probably the overall most flexible and powerful environment for processing large amounts of data (covering scenarios that cannot be fulfilled by MongoDB or Elastic Search).
Which one is the best then? It isn't possible to define an absolute winner: it depends by your real needs. In some scenarios you could also choose to work with a combination of more than one of them (there are existing use cases with Elastic Search + MongoDB or Hadoop + Elastic Search).
For the purposes of my current scenario I decided upon MongoDB, but in future scenarios the choice could be different.
- ElasticSearch
- MongoDB
- Hadoop
Lets' have a quick look at the features of each one of them first.
Elastic Search
Elastic Search (http://www.elasticsearch.org/) is a search engine based on Lucene (http://lucene.apache.org/). Even if some data visualization and analytics features have been added to it, it is basically a pure search full text search engine into schema free JSON documents.
MongoDB
MongoDB (http://www.mongodb.org/) is a cross-platform NoSQL document-oriented database. Data are stored in JSON-like documents into collection in dynamic schemas. The data interchange format of the documents is BSON (Binary JSON). BSON types are a superset of the JSON types. Like Elastic Search, MongoDB allows full text search in documents, but it provides more useful features like server-side scripts in JavaScript, MapReduce (a data processing paradigm for condensing large volumes of data into useful aggregated results) support, capped collections (the possibility to define a maximum size for a collection: so once it fills its allocated space, it makes room for new documents by overwriting the oldest documents in the collection; this feature could be useful capturing logs or any streaming data for analysis), and an aggregation pipeline framework.
Hadoop
Apache Hadoop (http://hadoop.apache.org/) is a framework for distributed storage and distributed processing of Big Data on clusters of commodity hardware. Its Hadoop Distributed File System (HDFS) splits files into large blocks (default is 64MB or 128MB) and distributes the blocks amongst the nodes in the cluster. For processing the data, the Hadoop Map/Reduce ships code (specifically Jar files) to the nodes that have the required data, and the nodes then process the data in parallel. At present time it provides probably the overall most flexible and powerful environment for processing large amounts of data (covering scenarios that cannot be fulfilled by MongoDB or Elastic Search).
Which one is the best then? It isn't possible to define an absolute winner: it depends by your real needs. In some scenarios you could also choose to work with a combination of more than one of them (there are existing use cases with Elastic Search + MongoDB or Hadoop + Elastic Search).
For the purposes of my current scenario I decided upon MongoDB, but in future scenarios the choice could be different.
Comments
Post a Comment