Drut Fabric Manager (DFM)

• Fleet of microservices deployed on a Kubernetes cluster.
• Managed via a user-friendly GUI within the Drut Software Platform.
• Exposes Redfish-compliant northbound REST APIs.
• Configurable for deployment per rack, group of racks, or entire clusters.

• Automatic Discovery: Identifies fabric elements and composable resources through OOB networks.
• System Composition: Creates systems with necessary data paths per configuration requirements.
• Dynamic Management: Attaches or detaches composable assets dynamically.
• Operational Support: Facilitates operations on composed systems.
• Health Monitoring: Conducts periodic checks on systems, data paths, and fabric elements.
• Event Management: Provides alarms and logs commands issued over the past 30 days.
• Firmware Upgrades: Automates updates for fabric elements.
• API Access: Features Role-Based Access Control (RBAC).
• Resilience: Independently scalable, fault-tolerant microservices with self-healing capabilities.
• Audit Trails: Tracks system operations, asset management, and metrics.
• Open-Source Foundation: Built using Spring Boot, Java, Python, RabbitMQ, PostgreSQL, and DockerCE.

Enterprise Data Centers

• Simplified scaling and reconfiguration of resources.
• Enhanced operational efficiency with dynamic fabric management.

High-Performance Computing (HPC)

• Seamless orchestration of diverse resources for compute-intensive tasks.
• Dynamic reallocation of assets to optimize performance.

AI Workloads

• Efficient management of GPU and IPU resources.
• Real-time adjustments to meet evolving AI demands.

Cloud Service Providers

• Streamlined creation of custom machine configurations.
• Enhanced fault tolerance and scalability.

Large Data Centers

• Centralized control for multi-rack and cluster-wide deployments.
• Reduced downtime with real-time resource reconfiguration.