Using
Elastic Search
as a complementary datastore
The
Problems
to SQL or to NoSQL ?
- Relational is good for:
- Storing complex structures
- Document-based is good for:
- Storing tons of similar object
- Scaling
to Store or to Compute ?
- Storing is good for:
- data safety
- speed (cache only : redis...)
- Computing is good for:
- save disk space
- speed (complex relations)
What do we
need then ?
A little bit of both !
Introducing ElasticSearch
- A datastore
- An index with a powerful set of API
- A REST service
- A very scalable product
- Used by the best companies out there (StackOverflow, Facebook, Goldman Sachs, NASA...)
What is Elasticsearch ?
Features
- Aggregated search
- Concept of relevance
- Partial/fuzzy matching
- Still a lot more to learn !!!
- As a complementary datastore
- As a search engine
- As a suggestion engine
How can we use it ?
Schemas speaks better
