INDUSTRY
Grids That Learn – Improving Application Efficiency with Moab
Cluster Resource, a leading provider of cluster, grid and utility computing software, announced today at GridWorld’s Exhibit Showcase the inclusion of Grid Learning in Moab Grid Suite -- a new feature designed to help maximize usage of grid resources. “Grid learning adds one more major feature to Moab’s abilities that help squeeze out every bit of a cluster or grid’s available resources and compute time,” said Josh Butikofer, Director of Grid Technologies at Cluster Resources.
Grid Learning provides added intelligence and adaptability to Moab Grid Suite - a policy based grid management solution that integrates scheduling, management, monitoring, and reporting of workloads across independent clusters. This intelligence enables Moab to learn strategies over time such as the allocation of optimal resources to a given application, the effective data transfer rate between clusters for automates date migration and expected job execution times for given users and accounts. Using this information, Moab continuously improves its workload management decisions, increasing system utilization and reducing average wait time. Moab Grid Suite ( www.clusterresources.com/mgs) is the first workload management solution to implement a feature such as Grid Learning into the scheduling decisions. Cluster Resources is further applying this technology to its Cluster and Utility Computing/On-Demand solutions, using the detailed knowledge of current environment and future decisions to make Moab’s peer-enabled services smarter. For example, Moab could enable a provisioning service that starts re-installing certain compute nodes with new images optimized for the changing workload. "Not only does Grid Learning further enhance Moab's ability to achieve 90-99 percent utilization of compute resources, but it also helps Moab pro-actively adapt to resource failures and environmental changes automatically," said Michael Jackson, President of Cluster Resources. "When you optimize your resources with a self-tuning system, jobs run faster and you accomplish more work in less time, resulting in improved return on investment across the grid." Grid administrators can tune Moab to adjust the impact level of Grid learning facilities allowing utilization, response time, fairness and other political factors to be appropriately balanced. Grid learning joins many features that help make Moab Grid Suite a leading grid workload management solution. Its task-based interface and flexible policy engine ensures that service levels are delivered, cluster autonomy is maintained and workload is processed faster. Moab optimizes data staging and integrates with existing security mechanisms or with grid security tool kits such as Globus. Moab Grid Suite decreases grid adoption barriers by giving sites extensive control over ownership-based access, priority levels and service levels to make the grid politically acceptable. It simplifies and unifies management across multiple clusters of heterogeneous hardware, storage, networks, licenses, resource managers and operating systems, by providing fine-tuned control over workloads and resources. Grid Learning will be available in Moab Grid Suite 4.5.2, which is in beta test now and scheduled for release in Q4. To download a free evaluation of Moab Grid Suite, visit its Web site.
Grid Learning provides added intelligence and adaptability to Moab Grid Suite - a policy based grid management solution that integrates scheduling, management, monitoring, and reporting of workloads across independent clusters. This intelligence enables Moab to learn strategies over time such as the allocation of optimal resources to a given application, the effective data transfer rate between clusters for automates date migration and expected job execution times for given users and accounts. Using this information, Moab continuously improves its workload management decisions, increasing system utilization and reducing average wait time. Moab Grid Suite ( www.clusterresources.com/mgs) is the first workload management solution to implement a feature such as Grid Learning into the scheduling decisions. Cluster Resources is further applying this technology to its Cluster and Utility Computing/On-Demand solutions, using the detailed knowledge of current environment and future decisions to make Moab’s peer-enabled services smarter. For example, Moab could enable a provisioning service that starts re-installing certain compute nodes with new images optimized for the changing workload. "Not only does Grid Learning further enhance Moab's ability to achieve 90-99 percent utilization of compute resources, but it also helps Moab pro-actively adapt to resource failures and environmental changes automatically," said Michael Jackson, President of Cluster Resources. "When you optimize your resources with a self-tuning system, jobs run faster and you accomplish more work in less time, resulting in improved return on investment across the grid." Grid administrators can tune Moab to adjust the impact level of Grid learning facilities allowing utilization, response time, fairness and other political factors to be appropriately balanced. Grid learning joins many features that help make Moab Grid Suite a leading grid workload management solution. Its task-based interface and flexible policy engine ensures that service levels are delivered, cluster autonomy is maintained and workload is processed faster. Moab optimizes data staging and integrates with existing security mechanisms or with grid security tool kits such as Globus. Moab Grid Suite decreases grid adoption barriers by giving sites extensive control over ownership-based access, priority levels and service levels to make the grid politically acceptable. It simplifies and unifies management across multiple clusters of heterogeneous hardware, storage, networks, licenses, resource managers and operating systems, by providing fine-tuned control over workloads and resources. Grid Learning will be available in Moab Grid Suite 4.5.2, which is in beta test now and scheduled for release in Q4. To download a free evaluation of Moab Grid Suite, visit its Web site.