Platform Computing now offers new solutions that allow you to dynamically provision operating systems (OS) and virtual machines (VM) in High Performance Computing (HPC) data centers and internal clouds.

Dynamic VM Provisioning

To increase utilization of valuable resources in an environment where jobs of many sizes and varied memory requirements are the norm, Dynamic VM Provisioning can be an effective solution. It reduces HPC management costs by dynamically provisioning virtual machines (VM) based on monitored workload and job resource requirements. Dynamic VM Provisioning, which leverages Platform LSF and Platform VM Orchestrator, is a unique solution offered only by Platform Computing. The result is an agile cluster optimized for maximum resource utilization and job throughput.

Dynamic OS Provisioning

If your organization operates a large HPC cluster that serves multiple departments or workgroups, it is very likely that many OSs will be required to run the applications each group uses. The challenge is to ensure that each group can run its application on demand, rather than wait for a server or node to be re-provisioned. Dynamic OS Provisioning reduces HPC workload management costs in heterogeneous clusters where fluctuating demand for various OSs makes it difficult to satisfy everyone’s needs with traditional, manual re-provisioning. This solution allows you to build a cluster in a configuration that best suits your needs, but retain the flexibility to provision resources to handle fluctuations in OS demand.

How you can benefit from Dynamic OS Provisioning and Dynamic VM Provisioning:

• Deliver higher service level to end users with a transparent, flexible and dynamic infrastructure 
• Save infrastructure and IT management costs with automated solution
• Increase SLA in a shared environment without sacrificing overall system utilization
• Provide historical workload data and resource requirement data for capacity planning

Selected facts about Dynamic OS Provisioning and Dynamic VM Provisioning:

• Enables dynamic system provisioning based on workload and policies
• Provides a centralized management console for manual provisioning and control of automated provisioning, and simplified troubleshooting and user support
• Isolates job run time environments from each other on multi-core systems
• Migrates jobs from one machine to another, without stopping them in order to optimize job scheduling and guarantee resource allocation and ownership
• Enhances fault tolerance, eliminating work lost due to hardware failure
• Optimizes application license usage by running a job on any slot, and later migrating it to the fastest machine