Managing scientific data for shared collaborative research has become a topic of growing importance. This is due in part to the increasing volume of data and technological advances that favor online, distributed, and massive data investigation. However, promoting such scientific practice brings challenges for processing, storage, and, interfacing with computational techniques. This last issue is particularly notable when final users are not computer scientists.
This paper presents SQLShare, a web-based system that abstracts the use of relational (conventional) databases with a SQL interface that is schema-free and transaction-free; accepts file-based incremental uploading; and supports metadata, tagging, provenance, and the simplified use of views, among other features.
Except for its interface, SQLShare uses two off-the-shelf solutions: Microsoft SQL Azure and Amazon’s cloud. Its features come from the functionalities afforded by these products, engineered as a back-end that manages SQL queries interpreted as views (materialized or not). The authors demonstrate the system with a massive biological oceanography project.
The authors are enthusiastic about their proposal, but not convincing about its contribution. Despite the lists of features and benefits, little is presented to support the authors’ claims. Besides, the paper presents too many details for biologists, and too few details for computer scientists. In sum, the paper is a nice piece of marketing, but falls a bit short as a scientific source, much like a white paper.