As vendor agnostic picture painters, we at Cutter have deployed many vendor cloud infrastructure platforms on top of various server hardware vendors integrating with a plethora of network vendors. Whether it was blue, green, red or purple hardware it didn’t really make any difference (nor should it) to the customer consuming the infrastructure as long as the build quality and support was in place, it mostly worked fine.
Dell (at the time owning VMware) brought to market a joint-engineered solution in VxRail sellotaping VMware’s VCF and Dell’s hardware together and offering it to customers of both. This was rather clever, taking two teams and having them sell the same product from two different angles. Other vendors did similar, working with VMware as the leading player to keep their hardware relevant alongside a customer needing cloud/server infrastructure.
Cisco however was there much, much earlier, though in a different way. Cisco UCS came to the market years earlier and allowed a Cisco network team retain control of the network, whilst building a bridge to a server platform team through a shared management plane that let them both enjoy their own slice of the Cisco pie.
Fast forward to Cisco Hyperflex, this was Cisco’s HCI solution supporting both VMware and Hyper-V, an evolution based upon the traditional UCS model whilst retaining the solid following it had already established.
Cisco and Nutanix with their many joint customers already, it made sense to bring out support for Nutanix on top of Cisco hardware and there has been growth in this area ever since, not to mention with the added pressures from Broadcom in 2024.
Cisco with their UCS Manager capabilities built into their Fabric Interconnect (FI) datacentre fabric product line provides a comprehensive management capability of the Cisco hardware and network layers, extended by Nutanix for the software layer on top is a well matched line up. Cisco whilst having UCS on-prem with FIs also have a cloud control plane product called Intersight which extends this management capability into the cloud, and further expands it to non UCS equipment.
Things get a little confusing at this point because depending upon which server hardware you pick, and whether you have FIs or TOR switches, and whether you want Intersight at the party or not leads a decision tree on ‘which management model(s) is/are available?’
Management Modes
Cisco has 3 management modes:
- UCS Managed Mode (UCS-M)
- Intersight Standalone Mode (ISM)
- Intersight Managed Mode (IMM)
Take for example, a common Cisco HCI UCS Nutanix node the ‘HCIAF240C M7 All-NVMe/All-Flash Server‘, if you have FI’s you are forced to use UCS Managed Mode (UCS-M)currently, which limits use of Intersight. This is further complicated because the FI’s must be in UCS Managed mode as well, not in Intersight Managed Mode (IMM).
If you use that same server with TOR switches you can only have Intersight Standalone Mode (ISM).
I hope Cisco bring Intersight Managed Mode (IMM) to the party for this rack mount server so we can have the best of both worlds, leverage FI’s (in IMM mode) and servers plugged into the FI so its all wrapped up in Intersight Managed Mode (IMM) as a whole.
If you go for one of the new Cisco UCS-X blade servers such as the ‘Cisco UCS X210c M7 Compute Node‘ these only support UCS Managed Mode (UCS-M) or Intersight Managed Mode (IMM) so need to be careful if you have both in your environment as you’ll have two different management modes.
Foundation or Foundation Central
Those of us used to building out Nutanix clusters will be familiar with the traditional Foundation process in taking nodes from the factory and bringing them into production. The traditional Foundation process is used to build UCS Managed (UCS-M) nodes into a cluster, once built, Nutanix LCM connects to the Cisco FI’s to coordinate firmware upgrades in the future by asking the FI/UCS Manager to perform firmware upgrades and return the node to Nutanix to remove it from software maintenance mode to carry on.
With Intersight Managed Mode (IMM) and Intersight Standalone Mode (ISM) we use Nutanix’s Foundation Central as part of Prism Central to connect to the Intersight service directly, obtain the onboarded servers and perform the cluster image and build process. This presents a little bit of a challenge in a greenfield environment where you may not have a Prism Central up front to perform the build. It is supported to deploy a single node Prism Central pc.2022 on a standalone ESXi/VMware Workstation/Hyper-V, upgrade Foundation Central and perform the build. Once built it is possible to then decommission the ‘temporary’ Prism Central instance and deploy out a new one using the one-click functionality. I hope this improves in the future to make the build process smoother, perhaps a cloud hosted Foundation Central ala Nutanix Central solution.
Once built, LCM activities connect to Intersight and Nutanix LCM coordinates all activities, much like what we’re used to with other OEM vendors.
Having built a fair few of these clusters in the recent months, I am impressed with the pace at which Cisco and Nutanix are expanding functionality and improving the process – there’s a lot of Cisco UCS and Cisco Hyperflex houses out there that could enjoy this partnership.