Cisco, Nutanix strengthen joint HCI package

Cisco and Nutanix have significantly expanded their alliance with new management capabilities, AI components and networking extensions for their integrated hyperconverged infrastructure (HCI) package aimed at easing edge, data center, and cloud operations.

Last year, Cisco killed its Hyperflex platform and essentially turned over that business to Nutanix with the idea that the vendors would work together to engineer collaborative technologies, services, support and sales. Following that decision, the vendors in October introduced Cisco Compute Hyperconverged with Nutanix. That package combines Cisco’s SaaS-managed compute and networking gear with Nutanix’s Cloud Platform, which includes Nutanix Cloud Infrastructure, Nutanix Cloud Manager, Nutanix Unified Storage, and Nutanix Desktop Services. The system can be centrally managed via Cisco’s cloud-based Intersight management platform, and it supports Nutanix Acropolis Hypervisor (AHV) and VMware vSphere hypervisors.

The vendors are now taking another step by announcing tighter integration between Intersight and Nutanix management systems; support for the Nutanix AI-based GPT-in-a-Box package; and improved enterprise network integration with Cisco’s Application Centric Infrastructure (ACI).

The management piece brings together the Cisco and Nutanix software management planes to give customers an integrated managed HCI architecture, which is available now, according to Jonathan Gorlin, director of product management, Cisco Compute group.

Today, Nutanix would integrate with servers directly, configuring them out of the box to run the Nutanix Cloud Platform (NCP) software stack. The vendors have now expanded that integration so that Nutanix Prism and Cisco Intersight management platforms work together to unlock Intersight’s full set of features within a Nutanix deployment, Gorlin said.

“The challenge that we see with our customers is trying to grapple with the complexity of infrastructure operations where they have structured and unstructured data proliferating at the cloud, edge, data center, you name it,” Gorlin said. “It’s becoming a very complex ecosystem to navigate and manage all those all those variables.”

By bringing the platforms together and extending the integration to the bare metal layer, it allows for a complete installation of a full software stack from scratch, completely remotely and with a fully automated workflow, Gorlin said. “For example, we can ship servers directly to an installation site, and a technician just needs to connect power network cables. You can perform all of the configuration, patching and deployment completely remotely,” he said.

“On top of that, when we combine Nutanix as flexible architecture, we can build clusters that start as small as one or two nodes and scale linearly to large data center footprints,” Gorlin said. The customer can monitor and manage the environment, see inventory, and control remote access, Gorlin said.

AI-ready infrastructure to simplify deployments

In the AI realm, Cisco has created a validated design, set to publish by end of May, for deploying Nutanix GPT-in-a-Box within the Cisco Compute Hyperconverged with Nutanix package. Nutanix describes GPT-in-a-Box as a turnkey package that simplifies the adoption of generative AI capabilities for enterprises. It provides an AI-ready infrastructure stack that allows organizations to deploy, run, and fine-tune large language models (LLM) and other AI workloads on-premises or at the edge, according to Ketan Shah, vice president of product management with Nutanix. 

“Deploying AI-ready infrastructure can be complex and costly,” Gorlin said. “GPT-in-a-Box is going to let our customers get up and running with all layers of the stack with verified, repeatable results,” Gorlin said. “Basically, we’re giving customers the recipe instead of clicking something from scratch; it’s always much easier to follow a recipe, and it will include support for a number of popular LLMs including llama and Falcon.”

Analysts said the Intersight integration and AI package should be significant for enterprise customers as they look to eliminate much of the complexity in managing and deploying new applications.

“Both the Intersight Standalone and GPT-in-a-Box are pretty significant for enterprise IT because they both remove a lot of friction. In the case of Intersight and Nutanix Prism, it’s about simplifying the deployment and provisioning – tasks that can be all-consuming for IT organizations,” said Matt Kimball, vice president and principal analyst with Moor Insights & Strategy covering data center compute and storage.

As for the for GPT-in-a-Box, it’s about simplifying the GenAI experience for enterprise IT, Kimball said. “I cannot tell you how many IT execs I’ve spoken with that talk about the challenges around sizing and deploying GenAI – from infrastructure to frameworks and LLMs to APIs to the management tools. This removes all of that complexity (and cost). And it addresses the top barrier to GenAI projects being successful,” Kimball said.

Networking integration

In addition to the Intersight integrations and AI support, Cisco and Nutanix have broadened their networking capabilities. 

For example, customers can now tie together systems via Cisco’s ACI package. ACI is Cisco’s intent-based networking technology, which gives customers the ability to implement network and policy changes on the fly and ensure data delivery. ACI Multi-Site typically lets two geographically dispersed data centers link via L2/L3 networks and offers consistent policy enforcement across both sites.

Customers can now integrate ACI and Nutanix AHV, the vendor’s hypervisor that lets customers to run multiple virtual machines on a single physical server. AHV is a core component of the Nutanix Acropolis Operating System (AOS), which provides the full software stack for Nutanix’s hyperconverged infrastructure platform, according to Shah.

Customers now can create application policies through ACI that push the network subnet configuration directly into Prism Central to be assigned to workloads. The integration also creates a security policy within Prism Central, the Nutanix cluster management system, that prevents communication of VMs with the associated category, tightening the security posture, Gorlin said. In addition, the integration provides unified visibility across compute and networking, making configuration and troubleshooting faster and in sync across IT teams, Gorlin said.

Also in the network realm, Gorlin said customers will have now have the option of using their existing fully integrated network with Cisco’s Unified Computing System (UCS) Fabric Interconnects or deploy their own network fabric.

UCS Fabric Interconnects act as a networking backbone for all Cisco UCS servers in combination with the Cisco Virtual Interface Adapter (VIC) to bring unified connectivity for LAN, SAN and management traffic over a single lossless Ethernet fabric, according to Cisco.

“UCS has always brought that high speed, east-west, high-performance networking fabric out of the box to support networking physical and virtual clusters, and that capability was available to the Nutanix clusters as well from Day 1,” Gorlin said. “Now we’re adding a secondary option to bring your own networking design to the solution. If customers have existing network fabric investments, or situations where they want to deeply customize their network design, this will allow you to kind of architect your own network. This means customers will have a choice for turnkey or customized networking depending on their requirements.”

Repurposing older servers for HCI use

The last piece of news features the ability for Cisco UCS customers to repurpose older UCS equipment for inclusion in the Cisco Compute Hyperconverged with Nutanix suite. 

“With Cisco UCS, we’ve pioneered the concept of a ‘software personality’ in our server baseboard management controller (BMC) that allows in-field reprogramming of servers to Nutanix HCI appliances,” Gorlin said. “The traditional scenario would require Nutanix HCI customers to purchase and deploy net-new factory-installed servers dedicated to running Nutanix.”

“Now, existing customers get to move freely to Nutanix without refreshing their hardware early, ultimately maximizing their investment and making operations more sustainable. The traditional scenario would require Nutanix HCI customers to purchase and deploy net-new factory-installed servers dedicated to running Nutanix,” Gorlin said.

Analysts said the Cisco-Nutanix partnership is focused on deeply engineering the technology between the two vendors rather than just reselling each other’s equipment, which will be valuable for enterprises looking at HCI.

“The integration is so different from what we typically see in terms of partnerships,” Kimball said. “This isn’t just about co-selling. It’s about delivering differentiation through integration. And I think it works for both companies,” he said. “Nutanix enables Cisco to come down market and sell into a segment that has typically shied away from them. And Cisco will bring Nutanix into the large enterprise market – a segment that has typically been very difficult for the company to penetrate.”