Verify our storage characteristics assumptions

Data storage is a sensitive topic. We offer our experiment recipes, environment descriptions, input data and expected output data to ensure that the code follows our ideas and that no regressions are introduced.

Scientific Data Repository

Our datasets are to be managed in a separate repository, following the best practices of scientific data management. It contains versioned performance measurements for the splitter (offline) and the data submission to the providers (online), including input data, generator scripts and descriptions. Check out the repository raocexperiments. (Warning: Due to the nature of extensive logging, this repository is quite large. Expect to download hundreds of MB.) The results are formatted and presented for comparison in Areca.

Storage Service Ontologies

We offer a CloudStorage domain ontology as well as specified instances thereof for existing real-world storage services and providers. The Web Services Modelling Language (WSML) is used to capture all technical aspects of the services such as its characteristics. Currently, the ontologies are maintained as just a set of files in a NubiSave repository directory.
We plan to run a permanently updated storage service directory with recommendations at some point in the future. For the time being, you will likely have to create your own provider description files unless you're happy with the defaults.