Broadcom’s VMware division rolled out an update to its vSphere Supervisor that makes it easier to update and manage Kubernetes-based workloads running in VMware’s virtualized environment, furthering VMware’s ongoing refinement of its vSphere Kubernetes Service (VKS).

The update involves unlocking the vSphere Supervisor platform from VMware’s vCenter so that customers can upgrade the former independently of their updates to the latter. Kishan Malur, technology product manager at Broadcom, explained in a blog post that this allows the “ability to deliver the Supervisor bits independently without using vCenter as a vehicle.”

The update also provides the Argo CD GitOps continuous delivery tool as a “Supervisor Service.” Argo CD is a Cloud Native Computing Foundation (CNCF)-based tool designed to help automate application deployment and lifecycle management.

VMware’s vSphere Supervisor allows users to deploy virtual machine (VM) or container-based workloads into vSphere using a declarative API. Malur noted this allows “cloud administrators to define resource boundaries, manage the platform with policies, governance, and standardization, while offering consistent self-service access to cloud services.”

Each vSphere Supervisor can support multiple VKS iterations so that the deployment of each VKS cluster can be coordinated with an approved set of Kubernetes releases. This provides users with some leeway in their use of Kubernetes releases that have been approved by their IT operations.

Broadcom’s bumpy Kubernetes journey

The VMware release update continues what has been some self-admitted jolting moves by Broadcom in refining its Kubernetes-focused operations.

Broadcom last year initially unlinked VKS, which was then known as the Tanzu Kubernetes Grid (TKG) from the Supervisor in a move that allowed for the updating of both products separate from the other.

Prashanth Shenoy, VP of cloud platform, infrastructure and solutions marketing at VMware, said in a press briefing at that time this allowed the embedded Kubernetes runtime to be “independently upgraded from the rest of the platform, so that you have that flexibility of when to get the latest and greatest from the Kubernetes runtime for their modern application independent of ESXi and other vCenter capabilities.”

“In the previous world they were not quite aligned to our existing vCenter updates and the vSphere release cycles that we have,” Shenoy said. “So what we have done is make sure that the TKG service can be aligned more with the upstream Kubernetes update so they can get the latest and greatest features independent of their vSphere release cycle.”

That move was quickly followed by the TKG-to-VKS name change, which VMware executives admitted did cause some confusion in the market. That confusion was heightened by the fact Broadcom continues to offer a stripped-down Tanzu platform alongside its bolstered VKS offering.

Tanzu is Broadcom’s pre-engineered platform-as-a-service (PaaS) designed to run on top of its halo VMware Cloud Foundation (VCF) or any infrastructure-as-a-service (IaaS) platform. It stands apart from VKS, which is a Kubernetes runtime and orchestration service that is included in VCF.

Broadcom has been steadily updating Tanzu since it closed on the VMware acquisition, including new generative AI (genAI) capabilities. However, Broadcom management has been more vocal in touting the benefits of VCF and vSphere offerings, which have overshadowed Tanzu.

Purnima Padmanabnan, GM of Broadcom’s Tanzu Division, during a press pre-briefing tied to last year’s Explore Barcelona event, towed that company line while touting Tanzu’s benefits.

“VCF and Tanzu platform are like peanut butter and jelly, because when you have a strong private cloud with VCF, and since we are all within the same organization, very, very tight integrations,” Padmanabhan said. “What you will see is, if you are on private cloud … don’t go for a bag of parts. Take VCF, do Tanzu platform and you have the full stack for application, for data, for any packages, for security, for operations, everything fully included, and that is the power.”

Broadcom, as usual, remains unfazed

Despite the drama, Shenoy earlier this year said that most of the vendor’s VCF customers were leveraging VKS to run their VMs and containerized applications.

“Any customer where 50% to 60% of their workloads are VMs – and the remaining are their cloud native or container applications – they're actively moving those workloads to leverage VKS because they've already paid for that as part of the platform, so they don't need to pay another platform,” Shenoy said.

Shenoy also explained that those containerized VCF workloads are coming from rival container-as-a-service (CaaS) platforms. The executive specifically pointing to Red Hat’s OpenShift as one of those drivers but added that customers going that route end up paying more.

Shenoy walked through an analysis of compute and cost requirements for an average customer running between 800 and 900 workloads in VMs and containers, which found a 50% cost savings for a customer relying exclusively on VKS.

“They have the flexibility and choice, but when we do the TCO analysis, it comes out more and more often that running natively using VKS, the container workloads is a lot more cost effective than having a two-platform strategy or running your containers on workload on bare metal, and that's where most of our customer conversations are,” Shenoy said.

That Red Hat comparison is significant as the IBM division has touted VMware conquest opportunities tied to Broadcom’s overhaul of VMware’s operations.

“What it’s forcing right now with the Broadcom acquisition is it’s forcing every enterprise client to make platform architecture decisions, and that’s going to be between virtualization and containerization,” IBM CFO James Kavanaugh said during an analyst conference last year, touting IBM’s consulting services, its Watson X genAI platform, and Red Hat OpenShift. “That’s why clients are coming to us with extreme interest around a growth factor that I think will play out for multiple years, and we’re excited about that.”

Andrew Sullivan, senior manager for virtualization and platform tech marketing at Red Hat’s Hybrid Platform Business Unit, furthered the notion of this Broadcom-driven VMware inflection point, telling SDxCentral in an interview that this has caused a pointed reaction from the market.

“My personal, anecdotal experiences this last year has been a lot of the emotional, visceral reactions to what’s happened, and I think people are digesting that and now they’re figuring out what those next steps are and not doing so in a panicked manner,” Sullivan said. “It’s less of the ‘oh, I need to move right now,’ and more of the, ‘OK, what’s my strategy going forward and how do I want to handle virtualization? Do I want to continue to have a single vendor strategy? How do I want to diversify across multiple hypervisors?’”

Sullivan added that those questions “very much triggered an out-of-cycle reevaluation.”

“I think folks are being more honest with themselves about what they need from a virtualization platform, and really giving a deep look into, ‘this was easy before and I didn’t have to invest much time, effort, energy, into it, it did everything I need. Should I continue doing that? How do I again diversify?’”

Red Hat’s claims are similar to those of other VMware rivals that have touted their own success in stealing disgruntled VMware customers away from Broadcom. However, Shenoy’s overall message continued what has been a common theme from Broadcom executives touting VMware’s post-Broadcom purchase success.