Hi, ElasticSearch (ES) and Jackrabbit (JR) cover different use cases, so can't really be compared with one another although there is some overlap.
ES provides search into documents, where the documents are bundles of properties represented by a json map. ES supports some complex query patterns that allow faceting and aggregation as well as pre and post filtering. The APIs to ES are ES specific and not to any standard. ES stores it data using Lucene segments and while many people have taken advantage of the scalability and redundancy of an ES cluster to use it as a primary persistence store, it was not really designed as such initially. JR is a content repository. It implements the JCR standard (JSR-170, JSR-283) with all the features that compliance to those standards require. It has well defined APIs (javax.jcr). It has strong persistence and a hierarchical structure to the content. The content may be versioned, observed, locked etc. The content may have a strong type model imposed on it. JR does have some query and search capability implementing JCR-SQL and JCR-XPath. Underneath JR uses Lucene for certain types of index. If you want a fully featured content repository, then JR is the right choice of the 2 systems. If you want an elastically scalable lucene based search system, then ES is the right choice. ES has more similarities to Solr than Jackrabbit. HTH Best Regards Ian On 9 September 2015 at 16:07, Adrian Luna <adrian.luna.co...@gmail.com> wrote: > Hi, > > I am new to Jackrabbit but have some experience with Search Engines and > lately with Elasticsearch. I can see the advantages of Elasticsearch over > Jackrabbit because of its distributed style. However, I think I don't > completely understand the picture and the space that Jackrabbit fills in > the world of technical solutions, because I can't figure out in which > situation could Jackrabbit be used instead of Elasticsearch. > Maybe because of my knowledge of Elasticsearch I tend to use it even > forcing it for systems where it's not designed for, so I would like to > understand the whole picture and maybe work with Jackrabbit to include it > into my skillset. > > I can imagine some of the issues could be support of a lot of document > types or observation, but still those are things that are implementable > using Elasticsearch maybe together with Apache Tika or simlar libraries. > > As I said, I just came up to Jackrabbit today, so my knowledge is very > limited, so don't be too harsh on me if the question is as simple as it > seems to be. >
