ApacheCon EU 2014

This is the story of building a clustered SPARQL database. Demand on
triplestores is growing, both in terms of scale and reliable operation.
Replication and clustering are needed for fault tolerant operation and for
larger datasets. Server replication of a single machine database only get so far.
It is easier to have a system designed to harness the capabilities of
several machines in one system. The goal of ""Lizard"" is to provide a solid
platform for triplestore operation.

In this presentation, Andy will describe how ""Lizard"" takes the Apache Jena
TDB database design and extends it to cluster operation. The talk will
cover the core principles of the TDB design and SPARQL execution, the
choices to be made in the ""Lizard"" design that rework the system for a
medium scale cluster system, and report of the current status.