Before listening, you might like to download the PDF that goes along with this show…
Building simplified, automated, and scalable data center networks with Cisco Unified Fabric is the topic of today’s show. Cisco Unified Fabric is an integrated end-to-end solution that builds on four fundamental pillars:
- Fabric management.
- Workload automation.
- Optimized networking.
- Virtual fabrics.
These features are applicable across the Nexus product line from the Nexus 7K down to the Nexus 1K with support currently available for the N55xx, N600x, N56xx, N2K, and N1KV. For users of Openstack and Cisco UCS Director, know that the physical network infrastructure is closely integrated with these orchestration engines.
A major part of the point of any fabric offering is to provide an end-to-end network solution for data center environments, which is exactly what Cisco’s Unified Fabric automation and management enhancements achieve.
On this Packet Pushers podcast, Cisco’s Shyam Kapadia and Lukas Krattiger discuss the following topics with co-hosts Ethan Banks and Greg Ferro:
- Spine-leaf based architecture with Layer-3 at the ToR switch.
- Overlay beginning at the ToR, allowing “lean” spines.
- Distributed anycast or pervasive gateway to localize the FHRP and broadcast domains.
- Multi-protocol BGP with L3-VPN address family for the control-plane.
- Segment ID and how is it used to overcome the 4K VLAN limitation, as well as to identify virtual fabrics.
- DCNM 7.x as the central point of management that packages various useful daemons such as LDAP, SCP, TFTP, and XCP.
- Automation of day-0 (device bootstrap and setup) and day-1 (Tenant) configs.
- Availability of Unified Fabric on existing Nexus devices you already own via a software upgrade.
Links
