#allegrograph

The main page of InfiniteGraph, a graph database commercialized by Objectivity, features an interesting comparison of 7 graph databases (InfiniteGraph, Neo4j, AllegroGraph, Titan, FlockDB, Dex, OrientDB) based on 16 criteria: licensing, source, scalability, graph model, schema model, API, query method, platforms, consistency, concurrency (distributed processing), partitioning, extensibility, visualizing tools, storage back end/persistency, language, backup/restore.

Unfortunately the image is almost unreadable, but Peter Karussell has extracted the data in a GoogleDoc spreadsheet embedded below.

Patrick Durusau mentioned on his blog a new record set by Franz’s AllegroGraph: 1 trillion RDF triples. This comes only 2 months after the previous Franz’s AllegroGraph record of 310 billion triples.

My first thought was: why is this important? It was one of the few times I’ve found the answer in the PR announcement:

A trillion RDF Statements […] is a primary interest for companies like Amdocs that use triples to represent real-time knowledge about telecom customers. Per-customer, Amdocs uses about 4,000 triples, so a large telecom like China Mobile would easily need 2 trillion triples to have detailed knowledge about each single customer.

The 310 billion triple result that Franz is announcing today was achieved in only two weeks of access (actual loading time of just over 78 hours) to an 8-socket Intel Xeon E7-8870 processor-based server system configured with 2 terabytes of physical memory and 22 terabytes of physical disk.

“We’re confident that with additional time, another terabyte of memory, and a bit more storage capacity, the previously unreachable goal of 1 trillion triples can be achieved. Even double that is not out of the question,” stated Dr. Jans Aasman, CEO of Franz Inc.

I’m afraid to ask how much would this cost. But we already know that scaling graph databases is still an open question.

This next answer shows why different data and processing models are needed for different scenarios:

Dr. Aasman said, “Some people have asked, ‘Why not do this on a distributed cloud system with Hadoop?’ The quick answer: NoSQL databases like Hadoop and Cassandra fail on joins. Big Enterprise, big web companies and big government intelligence organizations are all looking into big data to work with massive amounts of semi-unstructured data. They are finding that NoSQL databases are wonderful if one needs access to a single object in an ocean of billions of objects, however, they also find that the current NoSQL databases fall short if you need to run graph database operations that require many complicated joins. A typical example would be performing a social network analysis query on a large telecom call detail record database.”