ElasticSearch? YES, sure!2017-02-14T09:00:00.000Z 2017-02-14T09:00:00.000Z Recently we receive many requests from customers on technology ElasticSearch (ES) and similar tools. In some cases, the technology is not a mandatory condition for writing successful project. So how do you figure out whether ES is for your project?
First, we should to understand what the ES is.
This is a search engine, based on Lucene and developed in Java. Also you finde Lucene, Solr, SolrCloud etc tools. One thing for understanding - Solr provides Lucene capability in easy to use way and based on Apache Lucena, so anything, true for Solr is also true for Lucene.
Now you have to choose between two leaders ES and Solr. Both projects are very mature. Both have many features. Both are stable. Both are designed to perform similar tasks. And both are released under the Apache Software License.
So, have to choose? There is one difference - Solr has been more geared toward text search, ES is aiming to handle analytical types of queries, too, and such queries come at a price. If you use metrics and monitoring, with ES for you. Solr also exposes the key metrics, but nowhere near as many as ES. Regardless, having comprehensive monitoring and centralized logging tools like Sematext’s SPM Performance Monitoring and Logsene Log Management and Analytics — especially when they work seamlessly together like these two do — is essential if you want to have a handle on metrics and other operational data.
A few reasons of using ES over Solr:
- ES easier to scaling, provides using demanding larger clusters with more data and more nodes.
- ES, easier to get started with, without too much understanding of how things work, but it could be dangerous when data/cluster grows.
- ES is more dynamic – data can easily move around the cluster.
And some more pleasant trifles for dev’s:
- If work a lot with Node.js, API access is much easier in native JSON than it would be if you’ll convert to XML and back.
- If you use a lot of CouchDB, ES is a very good for that as its "river" mechanism plugs straight into CouchDB's replication stream.
- You can configure the schema over HTTP, which allows a bit more operation freedom
- ES takes a minimalist approach to configuration, which suits the style of the rest of my stack, i.e. dynamic types and schemaless documents
And give to you charts, for making your own desigen:
Solr user mailing list traffic (search-lucene)
Elasticsearch user mailing list traffic (search-lucene)