Verified Scalability Guide for Cisco Nexus Dashboard Fabric Controller 12.0.2f

Quick guide from the manual

This document provides verified scalability values for Cisco Nexus Dashboard Fabric Controller (NDFC) release 12.0.2f. These values are based on validated testbeds and represent practical operational limits rather than theoretical maximums. Key takeaways include support for 150 switches per fabric and 350 switches across the entire NDFC service.

Server Resource Requirements

To run NDFC on top of Nexus Dashboard (ND), specific hardware resources are required based on the deployment type:

• Virtual Node (vND) - app OVA: Requires 16 vCPUs, 64GB RAM, and 550GB SSD storage.
• Physical Node (pND) - SE-NODE-G2: Requires 2x 10-core 2.2G Intel Xeon Silver CPUs, 256GB RAM, 4x 2.4TB HDDs, 400GB SSD, and 1.2TB NVME drive.
• SAN Controller (Data Node vND): Requires 32 vCPUs, 128GB RAM, and 3TB SSD storage.

Scale Limits for Fabric Controller and Discovery

Scalability limits vary by deployment profile:

• 1-Node vND (app OVA): Up to 25 switches (Non-Production).
• 3-Node vND (app OVA): 80 switches.
• 3-Node pND (SE) / 5-Node vND (app OVA): 350 switches.

VXLAN EVPN (Greenfield) and Brownfield Migration

For new VXLAN EVPN fabrics, the limit is 150 switches per fabric. When transitioning existing fabrics (Brownfield Migration), the limit is 100 switches per fabric. Both scenarios support 50,000 endpoints and 150 networks with 4K IP allocations on the Infoblox server.

SAN Controller Limits

Limits for SAN Controller deployments depend on whether SAN Insights is enabled:

• Without SAN Insights: 80 switches and 20K ports for both 1-Node and 3-Node vND deployments.
• With SAN Insights: Ranges from 120K ITLs/ITNs (1-Node) up to 250K ITLs/ITNs (3-Node pND SE).