How does FlockDB compare with neo4j?

Both FlockDB and neo4j are open source frameworks for keeping large graph dataset. Anyone familiar enough开发者_运维百科 with both products to write a comparison?

---

For starters, here is what the FlockDB readme says:

FlockDB is much simpler than other graph databases such as neo4j because it tries to solve fewer problems. It scales horizontally and is designed for on-line, low-latency, high throughput environments such as web-sites.

As one example, FlockDB isn't capable of doing deep graph traversal because it's designed to handle Twitter's single-depth followers/following model.

The video here has a pretty good summary from the CEO of Neo Technology where he briefly mentions FlockDB: http://www.nosqldatabases.com/main/2010/7/5/need-a-graph-database-look-no-farther-than-neo4j.html

---

FlockDB and neo4j are quite different.

FlockDB provides distributed storage capabilities still trusting MySQL as backing storage; it is thought to query and create/update graph edges as fast as possible.

neo4j implements the full stack of storage services; it is thought to provide a comprehensive set of Graphs visit capabilities, even including a dedicated OO framework to easily write complex Graph visit algorithms.

A good introduction to FlockDB can be found here on Twitter Engineering blog.