Cisco Dynamic Fabric Automation Migration Guide
Migration Overview
Download this chapter
Migration OverviewMigration Overview
Download the complete book
Cisco Dynamic Fabric Automation Migration Guide Cisco Dynamic Fabric Automation Migration Guide (PDF - 5 MB)
 Feedback

Migration Overview

This chapter contains the following sections:

Prerequisites

To prepare for migration to the Cisco Dynamic Fabric Automation (DFA) solution, you must meet the following prerequisites.

  • Install and configure Cisco Data Center Network Manager 7.0
  • FabricPath on Spine-Leaf Topology
    • Nexus 7000 spine switches with NX-OS 6.2.(2) images
    • Nexus 6000 border leaf switches with NX-OS 6.02.N2 images
    • Nexus 6000 leaf switches with NX-OS 6.02.N2 images

      Note

      All non-Nexus 6000 boxes must be physically replaced with Nexus 6000 boxes with NX-OS 7.0(0)N1(1).

    • Nexus 1000v Series virtual switches at the virtual machine access layer

Existing FabricPath Topology

The existing FabricPath topolgy from which you are migrating includes:

  • An access layer with FabricPath-enabled VPCpath peers (VPC+)
  • Layer 3 aggregation layer-only connection to Spine layers
  • Two peers of Layer 3 boxes
  • Switched Virtual Interfaces (SVI) on only one set of VPC+ peers
  • HSRP running in local Layer 3 VLANS
Figure 1. Figure: Pre-migration Fabric Topology



Cisco Dynamic Fabric Automation Topology

An illustration of the Cisco Dynamic Fabric Automation (DFA) topology is shown in the following figure.

Figure 2. Cisco DFA topology



You can structure your Cisco Dynamic Fabric Automation (DFA) topology with two distinct fabrics:
  • Fabric with a mix of Nexus 5000 and Nexus 6000 leaves
  • Fabric with only Nexus 6000 leaves
Figure 3. DFA Fabric with a mix of Nexus 5000 and Nexus 6000 leaves



Figure 4. DFA Fabric with only Nexus 6000 leaves



The Cisco DFA fabric with both Nexus 5000 and 6000 leaves includes the following:

  • Nexus 5000 remains as Layer 2
  • Spine switches that can forward both 1q and 2q traffic, encapsulated in a FabricPath header
  • VLAN/SVI distinctions:
    • On a Nexus 5000, the VLAN/SVI is non-Segment ID-enabled across all Cisco DFA leaves running anycast gateway mode on Nexus 6000 leaves. Border leaf runs HSRP/VRRP as well as anycast gateway
    • On a Nexus 6000, the VLAN/SVI is Segment ID-enabled. The forwarding mode can be either proxy or anycast gateway.
    • Multicast will continue to run in the legacy multicast mode. Cisco DFA multicast should not be turned on.

The DFA fabric with only Nexus 6000 leaves includes the following:

  • Nexus 6000 leaves running either Anycast Gateway mode or Proxy Gateway mode
  • Spine switches that can forward both 1q and 2q traffic, encapsulated in a FabricPath header
  • VLANS that can be Segment ID-enabled

Traffic Flow Before and After Migration

As a result of changes to the topology and configuration of switches, traffic flow is optimized after the migration. Differences in traffic flow are shown in the following set of figures:

Figure 5. Figure: pre-migration inter-vlan single hop. Prior to migration, Inter-VLAN traffic from Host 1 on VLan10 goes through single Layer 3 hops up through the spine to get to host 2 on Vlan11.



Figure 6. Figure: post-migration inter-vlan single hop . After migration to the Cisco DFA fabric, inter-Vlan traffic from Host 1 on Vlan 10 takes a single hop through a single leaf node, where a Layer 3 lookup is performed and traffic is routed to host 2 on Vlan 11. Border Leafs start to respond to address resolution protocol (ARP) with anycast gateway media access control (MAC).



Figure 7. Figure: inter-vlan trafic multiple l3 hops x. Prior to migration, traffic going from host1 on vlan10 to host 5 on vlan20 takes multiple Layer 3 hops up to the Nexus 7000 Layer 3 and a series of Layer 3 lookups.



Figure 8. Figure: post-migration Unicast traffic flow . After migration, unicast traffic going from host 1 on vlan 10 to host 4 on vlan 20 takes fewer Layer 3 lookups at the leaf-level, and direct forwarding occurs between border leaf pairs through the spine without going to the Nexus 7000.



Figure 9. Figure: post-migration unicast traffic flow. Another illustration of post-migration unicast traffic flow.



Figure 10. Figure: North South Traffic Flow. North-South traffic remains unchanged after the migration and requires two Layer 3 lookups before reaching the Layer 3 cloud



Figure 11. Figure: Pre-migration Multicast Traffic Flow. PIM-SM and multicast replication behavior is the same as a non-FabricPath topology. Layer 2 multicast forwarding follows a pruned FabricPath tree. Internet Group Management Protocol (IGMP) is propagated to all FabricPath nodes via Intermediate-system to intermediate-system (ISIS).



Figure 12. Figure: post-migration multicast traffic flow . IF there is a Nexus 5000 in the topology, legacy multicast will continue to run.