About the CWS Virtual Desktop

About the CWS Virtual Desktop

The CWS Virtual Desktop (VDI) service has been developed to provide users interactive access to a virtual machine in the NCI HPC partner cloud. The VDI consists of a well configured compute node and dedicated CPUs, memory and local fast scratch SSD, and directly mounted access to the global lustre storage.  Users of the VDI can access an enhanced set of climate and weather tools that includes both the commonly used packages (such as those available on raijin), as well as new workflow and visualisation tools, including the climate workflows.

The VDI configuration includes:

  • Dedicated node on the NCI private cloud (initially a fixed size offer of 4CPU, 20GB RAM, 70 Gbytes SSD)
  • Direct access to data stored in the NCI global lustre filesystem (/g/data/) such as the CMIP5 data library.
  • Interactive tools such as ncview, Matlab, Python, Ferret, R. For licensed software, users can either provide their own standalone license or can connect to a remote license server.
  • All typical NetCDF tools are available (eg: NCO, CDO, and above)
  • VisTrails to streamline pipeline data processing, including access to the CWS Workflow Tool.
  • UV-CDAT (Ultrascale Visualization Climate Data Analysis Tools).

Connecting to the VDI

Follow the VDI setup instructions (Google Docs repository) to  connect via the “Strudel” client. Strudel was developed by Monash University and is available for Linux, Mac, Windows and requires TurboVNC. Users authenticate using NCI credentials (LDAP) and are then connected to their own VM on the NCI cloud. These nodes are managed by an NCI deployment of SLURM on the cloud that ensures that users have fast access to the dedicated node, and that network traffic is made directly between the VDI node and the user’s desktop.

VDI tools and resources

Once connected to the VDI, explore the Analyis and Visualisation tools on offer including the NCI installation of UV-CDAT and the CWS Workflow Tool.

Contact NCI helpdesk for more information.