Hacker News Reader

Comment by vosper | Hacker News Reader

I'm considering different graph DBs for a problem at work. I'd love to hear anyone's experience with Titan in production (or any other graph DB, for that matter!)

If you're in SF or Oakland and I could buy you lunch or a beer and talk about graph DB's with you for an hour, please find my email in my profile and drop me a line :)

View on HN

I can't speak about Titan, but Neo4j HQ is in San Mateo and I'm more than happy to buy anyone lunch or caffeine and discuss graph databases. Neo4j is not the right choice for every situation. But according to most metrics (e.g. http://db-engines.com/en/ranking/graph+dbms) it's around 8x more widely deployed than any other graph db, so we do have a lot of accumulated practical know-how on what works and doesn't when it comes to running graph databases in prod. I'm emil at neotechnology, just drop me a note if you're interested.

I tried really hard to use Titan for my project, but the tooling and the query language is —in my view— just atrocious. Setting up a running instance, trying to import my moderately complex property graph data into it, and navigating Tinkerpop's "funny" documentation was a real pain.

Then I decided to switch to Neo4j, and I was up and running in literally an afternoon. Also, its Java API is very well designed, which allowed me to fork and extend an integration plugin with Elastic in a couple days (https://github.com/jazzido/neo4j-elasticsearch/)

I've used a bunch of graph databases and I find they all fall short in one way or another. Almost all of them don't do anything that a good relational model can't do (except maybe faster for some queries). The biggest thing that leads me to believe that the graph database arena is not mature is that Tinkerpop is becoming the de facto standard. Talk about a joke. It is one of the silliest frameworks ever designed. Most of it's promise is just promise. Apache taking it over will probably lead to it's death.

Anyway, I'm going to give titan a second look with dynamo. I'm using OrientDB and neo4j right now. Both have major scaling issues, we haven't reached their limits yet but expect too very soon. I am wary of Titan because of Tinkerpop.

Can you elaborate on what scaling issues you've run into with Neo? I'm using it for my thesis project so I don't have very strict scalability requirements, but would love to know.

I second this, I'd love to hear about Titan in a production context, and I'm also in SF - would welcome an email.

Reply on news.ycombinator.com