* The graph is represented through bitmap data structures that allow high compression rates. * Node adjacencies are represented by bitmaps to minimize their footprint. * Each value in the database is represented only once, avoiding unnecessary replication. * Each of the bitmaps is partitioned into chunks that fit into disk pages to improve I/O locality. * The number of times each data page is brought to memory is minimized with advanced I/O policies. * It provides direct access to OS buffers, avoiding the decoding and encoding of other solutions using bitmaps, operations are computed with binary logic instructions that simplify the execution in pipelined processors. * The C++ core avoids overhead execution and complex memory management, as opposed to Java based engines. * High concurrency and low latency query response times secure high performance even under stress conditions. Full native indexing allows an extremely fast access to each of the graph data structures.

* Elastic and linear scalability for a growing data and user base. * Data distribution and replication for performance and fault tolerance. * Multi-datacenter high availability and hot backups. * Support for ACID and eventual consistency. * Support for various storage backends: - Apache Cassandra - Apache HBase - Oracle BerkeleyDB * Support for global graph data analytics, reporting, and ETL * Support for geo, numeric range, and full-text search via ElasticSearch, Solr, and Lucene * Native integration with the TinkerPop graph stack