Elephant - Persistent Key-Value store with Full Text Search | Hacker News Reader

Elephant - Persistent Key-Value store with Full Text Search

github.com

Seems cool, but the name should be changed. There is already an open source tool called ElephantDB created my Nathan Marz (Clojure, Big Data, Hadoop guru)

Not trying to be a jerk, I'm just sure that if this were my creation I would want people to let me know if I was potentially walking into brand confusion.

There is also John McCarthy's database-like language, Elephant 2000. Elephants never forget, so I think there will be many future databases named Elephant.

There is also this, http://common-lisp.net/project/elephant/

Not a name, but PostgreSQL has long used an elephant logo.

Wah. I was hoping for something more interesting than S3 for persistence of the key / values...as kenneth seems to have very novel and interesting takes on things.

Unless I'm mistaken, it's using ElasticSearch, not S3. It's, effectively, a niche ElasticSearch API tailored to key/value storage.

Elastic for the searching, but S3 for the persistence it appears. AWS credentials are even "expected" per the readme.

Ah... thanks for the clarification.

This basically defines an API for key-value storage. You could actually write an adapter for MySQL using full text search and MyISAM tables. You wouldn't want to, but you get what I'm saying.

I was expecting something like a Solr or an ElasticSearch, but this is more an application that limits what you can get out of a more powerful backend. I don't see why I would want that other than to encourage devs to become experts on a key-value store instead of learning the more complex internals. That's pretty much shooting yourself in the foot in terms of extensibility and flexibility. Better to know ElasticSearch and use it as a KV store than the other way around.

Very neat idea. I was thinking of something like this to help me offload some of my content from MongoDB to S3 but was stuck on how to make an efficient search.

Very cool, would be great to outline why one would use something like as opposed to say Riak.

To run Riak in production you need 5 servers.

Here all you need it S3(scales indefinitely) + 1 elasticache instance.

Why 5? We have a riak cluster running of 3.

"Here at Basho we want to make sure that your Riak implementations are set up from the beginning to succeed. While you can use the Riak Fast Track to quickly set up a 3-node dev/test environment, we recommend that all production deployments use a minimum of 5 nodes, ensuring you benefit from the architectural principles that underpin Riak’s availability, fault-tolerance and scaling properties."

Awesome selling point.

There are things that are missing that are nice about Riak. Mainly libraries, map reduce, 2i, and solr-like text search.

This project definitely interesting for what it is though.

Riak of course has a lot more features

Neat and I love the name! Who said elephants can't be agile & fast?

Biologists? Physicists? I don't know, who?

As always Kenneth Reitz did an awesome job.

Reply on news.ycombinator.com