Distributed database for tree structured data?

 

I'm looking for a database that is designed to handle tree structured data. I've looked into Neo4j and it's nice, but it doesn't really support sharding. Couchbase on the other hand isn't really built for a single tree of data (from my understanding).

A tree of data is theoretically really easy to split between servers whereas a complete graph isn't. I don't need the benefits of a full graph as much as I need the ability to split the big dataset dynamically. The full dataset won't fit on a single server.

The functions i need are (explained as if it were a normal computer file system):

  • Select every node in some "folder" excluding "sub-folders"
  • Split database (and load) between servers with redundancy

    • The system load will be mostly reads (approximately 90%).

    Thanks!

    I commented because I wanted to try and help and didn't think my link was really answer worthy. =)

    I've been keeping an eye on Titan and it seems to be very bleeding edge right now. After reading the article (http://thinkaurelius.com/2013/03/30/titan-server-from-a-single-server-to-a-highly-available-cluster/) it seemed like they are using Cassandra in some capacity, so the scale of Cassandra is tightly bound to the scale of Titan.

    From the article it seems that you don't have to explicitly shard the data. It will automatically distribute it for you when it reaches a certain size.

    Here is another presentation about Titan and how it uses Cassandra: http://blog.andreamostosi.name/2013/03/titan-distributed-graph-database/

    链接地址: http://www.djcxy.com/p/35038.html

    上一篇: 数据库的N项容量

    下一篇: 树结构化数据的分布式数据库？