Adaptive Computing’s Moab HPC Workload and Resource Orchestration Platform, already a world leader in dynamically optimizing large-scale HPC computing environments, has been enhanced to automatically extend on-premise HPC workloads into public clouds. Moab is now fully integrated with HPC Cloud ODDC, which works with Moab to deliver cloud services, enabling true scalability and elasticity.
ODDC Connect can provision a full stack across multiple cloud environments, and can be customized to satisfy multiple use cases and scenarios. Access to unlimited HPC compute resources is available in the cloud from multiple cloud providers. For less than the cost of one employee, peak workload demands can be met rapidly and accurately.
All required workload resources are automatically deployed as needed, and then retired when the resources are no longer required. When the workload requirements have been met and the resources provisioned in the cloud are no longer needed, Moab releases the bursting nodes from the cluster, and automatically deprovisions the cloud resources from the cloud provider. This added flexibility enables admins to expand their own cluster and dynamically utilize the scalability of the cloud.
Is the value of achieving HPC results at a guaranteed response time important for any critical workloads?
If so, this can be the financial justification to add HPC cloud bursting to the environment. HPC organizations must consider both lost value from inaction and costs to decide if bursting to the cloud is more cost effective than going through a traditional purchase process. If a surge of workload requests will cause a mission-critical product to not make it to market on time, or a high-value grant program to miss its delivery target, then the service guarantee achieved with bursting can fully justify the cost of HPC cloud bursting.
HPC cloud bursting helps avoid a number of supplemental costs, including cooling, power and facility costs, as well as expensive personnel time involved in procurement, upgrading systems, and decommissioning all of the bursted cloud resources. Further, access is gained to unique, specialized resources that will drastically improve the performance of certain workloads without having to justify the acquisition of the fixed resources for those special needs.
Rather than missing financially critical deadlines, or missing SLAs, organizations can add resources without incurring many of the overhead costs associated with on-premise resource purchases.
- Feature-rich bursting capability
- Industry-leading cluster utilization
- Robust policy and SLA enforcement
- Highly customizable for different cluster configurations
- Flag jobs to run either on-premise or in the cloud
- No additional hardware required, resulting in huge savings
- Integrates with on-premise resources
- Cloud resources are automatically deprovisioned from the cloud provider when no longer needed
- Can burst to multiple cloud providers (AWS, Google Cloud, Azure, Oracle Cloud, etc.) and bare metal provisioning
- Bursting is based on backlog
- Unneeded cloud nodes can be cleaned up early
- Bursts can be done in blocks of nodes or based on highest priority job
- Cloud nodes are totally dynamic inside of ODDC Connect – no need to predefine them ahead of time
- Safety limits for bursting can be set on a daily, weekly, quarterly, and yearly basis
- Completely automated – no admin/user interaction
- Amazon Web Services (AWS)
- Microsoft Azure
- Google Cloud
- Oracle Cloud Infrastructure
- Open Telekom Cloud
- Any Cloud Provider
- VMware vSphere
- VMware Cloud Director