Evaluation of Data Storage in HathiTrust Research Center Using Cassandra

Abstract

As digital data sources grow in number and size, they pose an opportunity for computational investigation by means of text mining, NLP, and other text analysis techniques. The HathiTrust Re-search Center (HTRC) was recently established to provision for automated analytical techniques on the over 11 million digitized volumes (books) of the HathiTrust digital repository. The HTRC data store that hosts and provisions access to HathiTrust volumes needs to be efficient, fault-tolerant and large-scale. In this paper, we propose three schema designs of Cassandra NoSQL store to represent HathiTrust corpus and perform extensive performance evaluation using simulated workloads. The experimental results demonstrate that encapsulating the whole volume within a single row with regular columns delivers the best overall performance.