SHARD - A distributed triple-store built on top of Hadoop

From lotico
Jump to: navigation, search

Date

January 9, 2011

Slides

PDF
Session-level: intermediate-advanced
Session-type: technology-coding

We will run a webcast with Kurt Rohloff from BBN early in 2011. Please be advised that this will be an advanced session with a technology focus and discussion around source code and implementation specific details BoF style.

Kurt Rohloff is a scientist at BBN Technologies. He recently designed, built and released a proof of-concept distributed triple-store called SHARD.

SHARD is built on top of the Hadoop MapReduce implementation to provide distributed, scalable performance. The SHARD source is on sourceforge with a BSD open-source license and configured to run with Hadoop 0.18. Kurt has given talks about SHARD to the most recent SemTech and HadoopWorld workshops. There's a small community of people working with SHARD that he is trying to grow and foster.

Here are the relevant pages with the code and documentation if you're interested: