We are one year into the Institute of Museum and Library Services-funded DRAS-TIC Fedora project, which emerged from discussions at CNI Fall 2016. Having completed the startup phase, we have begun to explore the first set of research questions set forth in our proposal, which concern the practicalities and performance implications of a Linked Data Platform (LDP) server based on Apache Cassandra NoSQL database. We completed extensive interviews with key staff at institutional partners to gather their recent repository needs and limitations. These are documented on our project test bed site, where visitors will soon find detailed performance graphs alongside links to software configurations and code.
Having established requirements, baseline code, and setup of the test bed cluster, we begin the exploratory phase of the project, in which we will create and run performance tests that measure several LDP server implementations under a variety of representative workloads. As we uncover gaps and opportunities we will iterate over own code and configuration, then document the outcomes. We are excited to put our own code and other implementations through their paces and forge a way forward toward next-generation repositories.
http://dcicblog.umd.edu/dras-tic-fedora/
http://drastic-testbed.umd.edu/
https://github.com/UMD-DCIC/gatling-testbed
https://www.trellisldp.org/
https://github.com/ajs6f/trellis-cassandra
https://www.imls.gov/sites/default/files/grants/lg-71-17-0159-17/proposals/lg-71-17-0159-17-full-proposal-documents.pdf
