Is it safe to use TitanDB?

Question

After the acquisition of Aurelius DataStax, and since Titan 1.0.0 was released in September 2015 and has received very few commits since then, I wonder if TitanDB can be used in production. Can anyone comment?

Answer 1

We are launching a Titan graph on top of HBase containing about 30 TB of data, and many features are missing.

For example, must-have is the ability to perform OLAP operations on a chart, such as removing redundant vertices using Spark.

Although it seems that Tinkerpop does just that using SparkGraphComputer, it doesn’t work very well - implementing reading data from HBase using the Hadoop InputFormat is a mistake and many scripts are not processed (for example, a vertex that is connected to itself in a loop makes the code excite exception and terminate). In addition, the performance of partitions that analyze vertices from raw data is simply poor — there are many redundant distributions that are redundant and make everything slow.

If you have been planning a long schedule for a long time, I don’t think Titan is suitable - unless you intend to use your own code.

Answer 2

Obviously, TitanDB was forked and now JanusGraph , here's an article about it.

Answer 3

[EDITED on April 28, 2017, to note that Titan is indeed dead; JanusGraph is a Titan plug and very cool]

I know several groups that use Titan in production, including people at IBM and Amazon. It's safe? Absolutely, if you have employees who can support him. (Shameless plugin: my company will help you with this.) Will the development of Titan continue? The development of TinkerPop is still developing very fast, although the contribution to the core Titan now comes mainly from non-Aurelius programmers.

I would argue that you are a fan of the Titan approach because of your size or data speeds, you should look at the transition to the new DataStax database when it is released this year. Work on this is at a feverish pace from what I can say while talking with Aurelius ex-guys at various conferences. If you are doing your job in Titan 1.0 (or 1.1, which is nearby) using TinkerPop 3, then this should be a simple transition to DSE Graph. This is what we offer our customers who need to manage very large data loads.

Answer 4

At the moment, it is unsafe to use TitanDB, because the main maintainers dropped it, and the last commit was more than a year ago. The company was acquired by Datastax, and the team was transferred to work on the DSE Graph . Although, few guys run Titan 1.0 in production.

A natural alternative is JanusGraph , which is essentially a well-preserved plywood fork for the unofficially released Titan v1.1. Check out their Github repository.

https://github.com/JanusGraph/janusgraph