Networking NVIDIA UFM Enterprise User Manual

Main Functionality Modules

Module

Description

Fabric Dashboard

The UFM central dashboard provides a single fabric-wide status view. The dashboard shows fabric utilization status, performance metrics, fabric-wide events, and fabric health alerts.

The dashboard enables you to efficiently monitor the fabric from a single screen and serves as a starting point for event or metric exploration.

Fabric Segmentation (PKey Management)

In the PKey Management view you can define and configure the segmentation of the fabric by associating ports to specific defined PKeys. You can add, remove, or update the association of ports to the related PKeys and update the qos_parameters for the PKey (mtu, rate, service_level).

Fabric Discovery and Physical View

UFM discovers the devices on the fabric and populates the views with the discovered entities. In the physical view of the fabric, you can view the physical fabric topology, model the data center floor, and manage all the physical-oriented events.

Central Device Management

You can use UFM to centrally access switches and hosts and perform maintenance tasks such as firmware and software upgrade, shutdown, and restart.

Monitoring

UFM includes a granular monitoring engine that provides real-time access to switch and server data. The engine collects and aggregates fabric and device health, traffic information, and fabric utilization data, and turns the data into meaningful information.

Configuration

You can perform in-depth fabric configuration, such as routing algorithm selection and access credentials, from the Settings view.

The Event Policy Table, one of the major components of the Configuration view, enables you to define threshold-based alerts on a variety of counters and fabric events. The fabric administrator or recipient of the alerts can quickly identify potential errors and failures and act to solve them.

Fabric Health

The fabric health tab contains functions for fabric bring-up and ongoing fabric operations. It includes one-click fabric health status reporting, UFM Server reporting, database and log snapshots, and more.

Logging

The Logging view enables you to view detailed logs and alarms that are filtered and sorted by category, providing visibility into traffic and device events as well as into UFM server activity history.

High Availability

In the event of a failover, when the primary (active) UFM server goes down or is disconnected from the fabric, UFM's High Availability (HA) capability allows for a secondary (standby) UFM server to immediately and seamlessly take over fabric management tasks. Failovers are handled seamlessly and are transparent to both the user and the applications running in the fabric. UFM’s High Availability capability, when combined with NVIDIA's High Availability switching solutions, allows for non-disruptive operation of complex and demanding data center environments.

For information about the main UFM functionalities, see the following sections:

Last updated: