On-Demand Data Center

The On-Demand Data Center, developed by Adaptive Computing, is a workload and resource management software and service solution designed for data centers, cloud environments, and high-performance computing (HPC) facilities. It caters to large enterprises and organizations of all sizes, spanning a wide range of professions and industries. According to the vendor, this scalable cloud systems management solution is used by government agencies, academic institutions, financial institutions, oil and gas companies, manufacturing companies, high-tech manufacturing, aerospace engineering, defense, universities, government research labs, scientific research institutions, engineering firms, financial services companies, and manufacturing companies.

Key Features

Per-Application Power Optimization: According to the vendor, the software automates CPU clock frequency on a per-application basis, allowing users to set specific clock speeds for different types of jobs. This feature aims to reduce unnecessary power usage, minimize energy costs, and maintain application-optimized power caps.

Eliminate Excess Power Consumption: The Green Pool feature, as claimed by the vendor, ensures that online nodes are always available for new jobs by powering down or powering up nodes as needed. This feature aims to save energy while maintaining system responsiveness and reducing the carbon footprint.

Remote Visualization: According to the vendor, users can render and visualize files remotely, reducing file transfer time and enabling effective collaboration through shared sessions. This feature is intended to save time, increase productivity, and eliminate the need for purchasing high-end desktops for all users.

Reporting and Analytics: The software, as stated by the vendor, streams resource usage and workload data into custom reports and personalized dashboards. This feature aims to improve resource utilization, enhance capacity planning and resource alignment, and enable better decision-making and policy enforcement.

Nitro High Throughput: According to the vendor, this feature enables high-speed throughput on short computing jobs by scheduling once for a large set of jobs, reducing launch time. It is intended to support thousands to billions of tasks in high-throughput computing (HTC) or service-oriented architecture (SOA) workloads, ultimately improving overall system efficiency.

Torque Resource Manager: The fully developed version of the TORQUE Resource Manager, integrated with the Moab HPC Suite, is customizable to specific system situations and provides control over batch jobs and distributed computing resources, according to the vendor.

Speed and Scale: According to the vendor, the Moab HPC Suite has been used on some of the largest computing environments globally. It supports multi-cluster/grid environments to balance workload management, accelerates launch time for large volumes of small jobs with Nitro High Throughput Manager, optimizes heterogeneous clusters with advanced resource management, and improves application performance with NUMA-aware job placement and GPU/Accelerator scaling.

Cloud Bursting: The software, as claimed by the vendor, enables on-premise systems to "burst" workload backlog to an external public cloud when resources are insufficient to accommodate peaks in demand. It automatically deploys and deprovisions all required workload resources as needed, allowing admins to expand their own cluster and dynamically utilize the scalability of the cloud.

Deliver Hybrid IT or a Pure Cloud Solution: According to the vendor, the software balances workloads between on-premise and cloud infrastructures, providing a true "hybrid cloud" experience. It eliminates long wait times by bursting workload backlog to the cloud, allowing for running multiple application types in the cloud, including new requirements like AI.

Auto-Deploy CI/CD Pipelines: The vendor states that this feature improves continuous integration/continuous deployment (CI/CD) by enabling automation at any part of the pipeline. It is designed to handle new pipelines quickly and deploy different toolchain combinations using the On-Demand Data Center (ODDC) framework.