One To Watch in the Big Data field
One To Watch: Aurelius's distributed graph database, Titan
We always like to profile some exciting projects that come over the horizon making big noises. One interesting project that has recently been released by graph engineering speciallists Aurelius is Titan - an open source, Apache2-licensed graph database tailored for Big Data problem solving.
Having expertise in the world of Neo4j, TinkerPop and R Statistics, Aurelius have offered their own alternative. The highly scalable distributed graph database leverages existing distributed storage engines such as Cassandra and HBase to facilitate the representation and processing of massive-scale graphs within a multi-machine cluster. A shrewd move, given how the multi-cluster approach is getting increasingly important for many Big Data developers.
The single machine has outstayed its welcome - it can no longer offer the storage and computational capacities that larger data graphs need. Enter Titan - it supports an 'infinite' sized graph and an 'unlimited' number of concurrent transactions through the addition of machines to the cluster
In addition, Titan natively implements the TinkerPop Blueprints API to ease the transition into existing architectures, supports graph traversal language Gremlin and integrates seamlessly with the Rexster graph server.
The presentation from Aurelius below explores the changing needs of developers towards graph databases, and how Titan can change things for the area. It's encouraging to note the support for already-established option Apache Cassandra and HBase, right from the early days.
It's a very impressive greeting from Titan - the team behind the project clearly know their onions, and can visualise them too! Titan looks ready to lift Big Data graphs to the pedestal they belong on, focusing on concurrency, scalability and being dynamic. Not to mention the inclusion of some important standards.
The future looks well planned out too, with the inclusion of Faunus, a path algebra for Hadoop, and Fulgora, an in-memory graph engine. Now they just need to shout about it, to gain adoption.
If you're intrigued by this latest NoSQL/Big Data contenders, check out the rather flashy GitHub home for it. Another plus point is the excellent documentation - overflowing with theory . That can only be a good thing. Keep an eye on this one as it appears fully in July. The Graph Database market is growing and we now welcome another to the fold.