Run data analysis in the CloudStor interface using CloudStor SWAN (Service for Web-based Analysis). You can run data cleaning and transformation, numerical simulation, statistical modelling, data visualisation, and machine learning processes, and share these with other SWAN users.
Access CloudStor SWAN through the SWAN tab at the top of the CloudStor web interface.
Data backup
Data in your SWAN home folder is regularly backed up to your CloudStor account.
Learn how to use SWAN and access backups
Jupyter Notebooks
Jupyter Notebooks are open-source electronic versions of a laboratory notebook. Through CloudStor SWAN using Jupyter Notebooks, users can create, execute and share documents containing code, equations, visualisations and narrative text through their web browser.
Jupyter Notebooks are available in eight different programming languages: R, Python, Octave, C++, Julia, Java, OpenRefine and RStudio. Additional code libraries in these languages can be imported and installed in a Notebook to meet different research needs.
You can learn more about Jupyter Notebooks on the Jupyter website.
Hardware resources
SWAN resources are shared amongst its users, and consists of a 14 node cluster connected to the AARNet backbone at 100 Gbps. Each SWAN node provides:
- x72 2.30 GHz CPU cores
- 256 GB RAM
- 25 Gbps active-backup network
- 5.5 TB of scratch space
Note: each SWAN session is limited to a maximum of 10 CPU cores and 16 GB RAM.
Processing capacity
SWAN is suited to moderate and exploratory data analysis of CloudStor data, and for sharing and collaborating through the Jupyter Lab environment. Researchers running large scale or long asynchronous analyses are encouraged to discuss large processing requirements with their institution.
CloudStor SWAN should only be used to run workloads appropriate to and within the guidelines of your institution. Abuse of the service will be logged and reported.