New and Changed Information

The following table provides an overview of the significant changes to the organization and features in this guide from the time the guide was first published to the latest update.

Table 1. Latest Updates

Date

Changes

January 25, 2021

First release of this document.

Overview

This guide contains the maximum verified scalability limits for Cisco Multi-Site.

These values are based on a profile where each feature was scaled to the numbers specified in the tables. These numbers do not represent the theoretically possible scale.


Note

The total number of objects within each site must not exceed the maximum verified scalability limit for that fabric version. For more information on site-specific scalability limits, see the Cisco ACI Verified Scalability Guide or Cisco DCNM Verified Scalability Guide for that fabric version.

ACI Fabrics Scalability Limits

This release supports managing only DCNM fabrics or only ACI fabrics by the same Multi-Site Orchestrator. The following scale limits apply when managing ACI fabrics.

General Scalability Limits

Object

Scale

Sites

12

Pods per site

12

Leaf switches per site

400 in a single pod

500 across all pods in Multi-Pod fabrics

Total leaf switches across all sites

Sites * Leaf switches per site

For example, 6000 if every site is deployed as a Multi-Pod fabric.

Endpoints

150,000 including:

  • 100,000 - learned from other sites

  • 50,000 - locally learned in site-local

Multi-Site Orchestrator Objects Scale

Object

Scale

Number of Schemas

80

Templates per Schema

10

Application Profiles per Schema

200

Policy Objects per Schema

1000

Contract Preferred Group (BD/EPG combinations)

500

This value represents the number of EPGs that are part of the Preferred Group (across all the defined VRFs) that can be deployed in each site. This means that the maximum number of EPGs in the Preferred Group that can be managed by a single Nexus Dashboard Orchestrator instance can range from 500 (if all the EPGs are stretched) to 500*12 if only site-local EPGs are defined in each site.

MSO-Deployed Objects Scale

To better understand the scalability values captured in the following table, it is important to clarify that there are three kind of MSO-deployed objects:

  • Site local objects—these are the objects defined in templates associated to a single site, which get deployed by MSO only in that specific site.

  • Shadow objects:—these are the objects deployed by MSO in a site as a result of a contract established between site local and remote objects, they are the representation ("shadow)" of the remote object in the local site.

  • Stretched objects—these are the objects defined in templates that are associated to multiple sites, which get deployed by MSO concurrently on all those sites.

The table below captures the maximum number of objects that MSO can deploy in a given site and includes the sum of all three kinds of objects described above.

For example, if you have two sites and you define three templates on MSO—template-1 associated to site-1, template-2 associated to site-2, and template-stretched associated to both site-1 and site-2—then:

  • If you configure and deploy EPG-1 in template-1, this will count as one EPG towards maximum allowed for site-1.

  • If you configure and deploy EPG-2 in template-2, this will count as one EPG towards maximum allowed for site-2.

  • If you apply a contract between EPG-1 and EPG-2 or add both EPGs to the Preferred Group), a shadow EPG-2 will be created in site-1 and a shadow EPG-1 in site-2. As a result, two EPGs will now be counted towards maximum allowed in each site.

  • Finally, if you configure and deploy EPG-3 in template-stretched, it will count as another EPG in each site, bringing the total to 3 EPGs towards maximum allowed scale.

It is worth adding that the maximum number of objects supported in a given fabric (and captured in the Verified Scalability Guide for Cisco APIC) must not exceed the sum of objects locally defined on APIC plus the objects pushed from MSO to that site (MSO-deployed objects).


Note

For maximum scale Multi-Site configurations with many features enabled simultaneously, we recommend that those configurations be tested in a lab before deployment.

Object

Scale (Stretched)

Tenants

400

VRFs

1000

BDs

4000

Contracts

4000

EPGs

4000

Isolated EPGs

400

Microsegment EPGs

400

L3Out external EPGs

500

Subnets

8000

Number of L4-L7 logical devices

400

Number of graph instances

250

Number of device clusters per tenant

10

Number of graph instances per device cluster

125

VRF/BD VNID Translation Scale

Object

Scale

Fixed spines

21,000

Modular spines

42,000

DCNM Fabrics Scalability Limits

This release of Multi-Site Orchestrator supports managing only DCNM fabrics or only ACI fabrics by the same Multi-Site Orchestrator. The following scale limits apply when managing DCNM fabrics.

General Scalability Limits

Object

Scale

Sites

6

Leaf switches per site

150 per DCNM fabrics

350 per DCNM instance

900 total

Border Gateways per site

4

Multi-Site Orchestrator Objects Scale

Object

Scale

Policy Objects per Schema

1000

Templates per Schema

10

Number of Schemas

80

Multi-Site Orchestrator Users (nonparallel*)

*Multi-Site Orchestrator processes requests sequentially from multiple users even if they are deploying different schemas.

50

MSO-Deployed Objects Scale

When MSO manages DCNM fabrics, there is no concept of "shadow" objects. Hence, the scalability values captured in the table below only refer to the sum of site-local and stretched objects deployed by MSO in a given site.

Object

Scale per Site

VRFs

500

Networks

1000 (L3)

1500 (L2)