ホーム / Cluster computing / Platform Cluster Manager / Workload Management

Workload Management

Platform Cluster Manager

Workload Management

Platform Cluster Manager
Workload management is a critical aspect of day to day cluster operations. With Platform Cluster Manager, workload management is more than just an after-thought – it is an integral part of the cluster management solution.
    • Free, open-source Platform Lava workload management
    • Pre-configured, immediately ready for use
    • An intuitive end-user friendly job-submission portal
    • Cluster changes automatically reflected in workload management configuration
    • Easy to use web-based job monitoring and management
    • Integrated reporting via the Platform Management Console
    • 100% command line compatible with Platform LSF for maximum investment protection
    • A range of available commercial add-on solutions for maximum flexibility
Built-in, ready for use:

Workload management in Platform Cluster Manager is tightly integrated and well designed. Advanced users may choose to interact with a well documented command line interface, however new cluster users will appreciate that jobs can be monitored and managed entirely through the web interface. When Platform Cluster Manager is installed, the Platform Lava (or optional Platform LSF workload manager) can be automatically included, configured, and made ready for use with a reasonable set of queues and scheduling policies. This makes workload management available for use immediately. All that is required of the administrator is to give permissions to cluster users, through the management console, allowing them to submit jobs to the cluster.

Web management and application portals:

Features like application aware job submission, job monitoring, management and reporting are all included in the workload management software. As you add or remove hosts, or change the software stacks deployed to different types of nodes, the cluster configuration files change automatically, providing a single point of administration. Competing solutions often require that the cluster manager and workload manager be separately configured resulting in additional work for busy administrators and an increased opportunity for error. Platform Cluster Manager, like any open toolkit, allows cluster administrators the freedom to deploy their choice of open-source or commercial workload management solutions. Most customers will appreciate the ease of installation, the ease of management, the compatibility and the extensibility offered by the direct integration of Platform Lava and Platform LSF with Platform Cluster Manager.

The Platform Management Console (PMC) provides standard job submission templates for a variety of popular HPC applications including ANSYS, ABAQUS, NASTRAN, FLUENT and LS-DYNA. Similar interfaces can easily be built using these templates for your own applications or other third party simulation tools. Customers requiring more extensibility or customization can integrate third party portal solutions offering enhanced functionality such as EnginFrame™, a popular portal solution for HPC clusters.

Expansion built-in:

Users who choose to start with the open-source Platform Lava kit included in Platform Cluster Manager can be assured of a trouble-free upgrade path. Platform Lava is a 100% compatible subset of Platform LSF, meaning that if an upgrade is required in the future, investments in scripts and user training is 100% preserved. Users can migrate to Platform LSF and even continue to use the same web-based job submission and management tools that they used with Platform Lava.

For users who have more sophisticated workload management requirements, or who are running parallel workloads, Platform LSF is recommended as a preferred solution. Customers running Platform LSF have a clear growth path since they can easily add layered applications at a later time including:

  • Platform LSF MultiCluster
  • Platform Analytics
  • Platform LSF License Scheduler
  • Platform LSF Process Manager
  • Platform LSF Session Scheduler
  • Platform MPI

To learn more

 

  RES - Renewable Energy Systems

 

;