Shay banon, creator and founder of elasticsearch give us a nice take at Zenika office.
Opportunity for me te review many things and confirm presentations are a complementary and to my advice a great boost for your learning curve.
- talk about sharding and replication basics
- how requests are routed to shards and replicas (essential to understand system performance)
- use aliases on index
- some usecases and common patterns to desing your system
- and many good questions et the end of the talk
Shay launch some nice troll i’ve really enjoyed:
- competitors ? who ? : yes, i dont know them too, he should be right
- nice info on poorly implemented http client. http is ok for perf…. but implementation in ruby, python and so on….
- on key value get, es is fastest than a lot of other key value search…. ;) like this one (index and proper caching power!!)
what i think
This tool is really great. When you see all possible applications (more than search ;) ) you could tend to use it as your main datastore but…
- es is for search, high availability, so think about consistency….
- it an index. high write cost, super fast reads. putting data in it could take a while….
I’m going to use it for a dataviz on open data. Super handy: grab jsons and push, query in realtime, have facets… and histograms !!!
- I like a lot the automatic mapping to types (dates, timestamps and so on….)
- For low writes / read only data its perfect.
- I like a lot the automatic mapping to types (dates, timestamps and so on….).