Graph Databases - What's So Different

Darren Wood, the Chief Architect at InfiniteGraph has a presentation that discusses what makes graph databases unique to their relational counterparts.

One very interesting concept discussed in this presentation is partitioning of graph data. Partitioning of graphs is a very complex and difficult problem namely because the graph is a living entity that is constantly changing i.e. new relationships between nodes.

Wood notes that there are critical must haves for distributed graphs:

• High performance distributed persistence
• Ability to deal with remote data reads (fast)
• Intelligent local cache of subgraphs
• Distributed navigation processing
• Distributed, multi-source concurrent ingest
• Write modes supporting both strict and eventually consistency