UFM Benefits

NVIDIA UFM Enterprise User Manual v6.15.0

UFM provides all fabric management functions in one central console.

The ability to monitor, troubleshoot, configure and optimize all fabric aspects is available via one interface. UFM’s central dashboard provides a one-view fabric-wide status view.

UFM includes an advanced granular monitoring engine that provides real-time access to switch and host data, enabling cluster-wide monitoring of fabric health and performance, real-time identification of fabric-related errors and failures, quick problem resolution via granular threshold-based alerts and a fabric utilization dashboard.

Advanced Traffic Analysis

Fabric congestion is difficult to detect when using traditional management tools, resulting in unnoticed congestion and fabric under-utilization. UFM’s unique traffic map quickly identifies traffic trends, traffic bottlenecks, and congestion events spreading over the fabric, which enables the administrator to identify and resolve problems promptly and accurately.

Consolidating multiple clusters into a single environment with multi-tenant data centers and heterogeneous application landscapes requires specific policies for the different parts of the fabric. UFM enables segmentation of the fabric into isolated partitions, increasing traffic security and application performance.

UFM uses a logical fabric model to manage the fabric as a set of business-related entities, such as time critical applications or services. The logical fabric model enables fabric monitoring and performance optimization on the application level rather than just at the individual port or device level. Managing the fabric using the logical fabric model provides improved visibility into fabric performance and potential bottlenecks, improved performance due to application-centric optimizations, quicker troubleshooting and higher fabric utilization.

UFM provides comprehensive information from switches and hosts, showing errors and traffic issues such as congestion. The information is presented in a concise manner over a unified dashboard and configurable monitoring sessions. The monitored data can be correlated per job and customer, and threshold-based alarms can be set.

Failovers are handled seamlessly and are transparent to both the user and the applications running on the fabric, significantly lowering downtime. The seamless failover makes UFM in conjunction with other Mellanox products, a robust, production-ready solution for the most demanding data center environments.

UFM provides an advanced Web Service interface and CLI that integrate with external management tools. The combination enables data center administrators to consolidate management dashboards while flawlessly sharing information among the various management applications, synchronizing overall resource scheduling, and simplifying provisioning and administration.

© Copyright 2023, NVIDIA. Last updated on Nov 8, 2023.