The Introduction of the OSCAR Database API (ODA)

The OSCAR [14] cluster installation toolkit was created by the Open Cluster Group (OCG) for one particular type of High Performance Computing (HPC) cluster. OSCAR is currently one of the widely used cluster installation toolkits; it boasts hundreds of thousands of downloads and active mailing lists. OSCAR has expanded its area with several sub-projects targeting other types of HPC clusters. Each of these projects share a core set of OSCAR code, including the OSCAR Database and its access API, "ODA" (OSCAR Database API). The ODA abstraction layer, consisting of a database schema and corresponding API, hides a commodity back-end database (e.g., MySQL [15]). Because OSCAR and its sub-projects are targeted at new, innovative environments (including non-HPC environments), there are significant issues with managing various configurations of each project. For example, as we previously showed [8], “previous versions of ODA were unable to represent the complex, ever-growing set of data required to accurately describe the clusters that it manages. Further, its API was extremely complex, requiring a steep learning curve for OSCAR developers”. Therefore, we have designed and implemented a new database schema to deal with these issues. This new version of ODA has not only resolved the above problems but also, as proposed in our previous paper, enabled storage and retrieval of various configuration information, and encouraged data re-use between the main OSCAR project and its derivative projects. In addition, the new version of ODA has sped up the OSCAR installation process. This document presents a simpler, highly flexible design and implementation of ODA slated to be included in OSCAR v5.0. It also suggests a blueprint for maintaining the database modules of ODA in a systematic, organized way.