VMware, Tanzu, and the Broadcom Reformation (2024–2025 Retrospective)
Ghosts of vClouds Past
I am an AI developer, Cloud Philosopher, Former Virtualization Evangelist, and Unofficial Archivist of Infrastructure Folklore - Follow my IT Archaeologist Podcast on YouTube.
Read my predictions for vCloud Director, back in 2011 - https://www.sanjaysays.com/2011/12/why-vmware-vcloud-director-wont-rule.html
Read my follow-up article on vCloud Director in 2016 - https://www.sanjaysays.com/2016/07/the-curious-case-of-vcloud-director-and.html
Let me take you back—no, not to 2011 this time, but to a tech universe that now feels almost ancient: a world where vCloud Director once promised every enterprise its own private AWS, and VMware was the emperor of the data center, wrapped in the silken robes of ESXi, vCenter, and vSAN.
But here we are now, in the twilight of 2024. vCD is an artifact, like a fossil embedded in the sediment of cloud history. VMware is no longer VMware, at least not in spirit. And Broadcom? Well, Broadcom did what Broadcom does best, optimize portfolios like hedge funds trim fat from spreadsheets.
So let’s talk about what happened, what it means, and why the legacy of vCloud Director is now just a cautionary tale whispered in DevOps Slack channels after midnight.
The End of an Era: Goodbye VMware, Hello Broadcom Stack Optimization Services™
When Broadcom completed its $61 billion acquisition of VMware in November 2023, we all knew changes were coming. What we didn’t expect, at least not with such surgical swiftness, was how decisively Broadcom would restructure, rebrand, and retire large swaths of VMware’s sprawling portfolio.
It took less than 12 months for Broadcom to:
-
Kill off or severely limit availability of VMware’s free products and SMB offerings
-
Retire the perpetual licensing model, moving everything to expensive, high-touch enterprise subscriptions
-
Refocus the sales motion on Fortune 500s and Tier 1 providers (read: “If you’re not spending eight figures, please see yourself out”)
But among the earliest victims of this rationalization? You guessed it: vCloud Director. Not just sunsetted, but quietly buried, with a dry press release and a PDF link to the Broadcom Lifecycle Matrix of Doom.
For those still running vCD in 2023, mostly aging service providers clinging to multi-tenant VM platforms, it was the final signal to pack up their custom catalogs, retire their fenced networks, and migrate to more modern platforms… or close up shop.
Tanzu: The Great VMware Reboot That Came Too Late
Now, let’s talk about Tanzu. VMware’s bold, ambitious, and ultimately tangled attempt to reinvent itself for the Kubernetes age.
Remember when Tanzu was introduced around 2019? It was positioned as the bridge from vSphere to Kubernetes, a lifeline for legacy VMware customers to leap into the cloud-native pool without cannonballing directly into AWS or GCP.
Tanzu had everything:
-
Tanzu Kubernetes Grid (TKG) vSphere-native Kubernetes clusters
-
Tanzu Mission Control Multi-cluster management with an elegant UI
-
Tanzu Application Platform (TAP) Dev-friendly abstraction layers
-
Tanzu Service Mesh, Build Service, and a buffet of Spring integrations
In theory, Tanzu was VMware’s redemption arc. A post-vCloud, container-native, DevOps-loving platform that could compete with OpenShift and integrate cleanly with cloud-native CI/CD pipelines.
In practice?
-
Tanzu was a branding mess. Product names changed more often than container images in a CI pipeline.
-
Dependencies were bloated. Some components required NSX-T, others didn’t. Installing Tanzu sometimes felt like assembling IKEA furniture blindfolded, using instructions written by committee.
-
Pricing was confusing, especially for smaller teams trying to scale incrementally.
VMware eventually streamlined Tanzu by 2022, but by then, Kubernetes adoption had matured, and enterprise developers were embracing lightweight distributions like K3s or cloud-managed services like EKS and GKE. Tanzu became a tool not of transformation, but of preservation, a way to bring containers to vSphere shops who couldn’t let go of their infrastructure past.
Then came Broadcom.
Broadcom’s Philosophy: Cut Deep, Monetize Heavy
Broadcom’s playbook is well-documented. Acquire mature enterprise software, slash R&D, focus on high-margin customers, and turn recurring revenue into Wall Street magic.
And Tanzu? It was too niche, too developer-focused, too… open.
By mid-2024, Tanzu Application Platform (TAP) was discontinued. Tanzu Community Edition had already been retired the year before. Broadcom reoriented Tanzu’s future into a tightly controlled enterprise-only SKU with deep integration into VMware Cloud Foundation (VCF), essentially, the new SDDC bundle.
If you were:
-
A bank with 400 developers, or
-
A service provider paying for NSX, vSAN, and VCF
…then Tanzu was still yours to play with.
But if you were:
-
A mid-market shop, or
-
A developer team trying to test Spring Boot on Tanzu, or
-
A vCD customer looking for a path forward
…then you were out of luck. No seat at the Tanzu table. No free lunch. Just a form on the Broadcom website labeled “Enterprise Kubernetes Licensing Request.”
The vCloud Director Legacy: A Case Study in Abstraction Overload
So now that the dust has settled, what is the legacy of vCloud Director?
Here’s what I believe it will be remembered for:
Too Much Control, Too Little Simplicity
vCD tried to give service providers an AWS-like abstraction layer, without the scale, elasticity, or economic model of any modern CSP like AWS, Azure, GCP, or OCI. It succeeded only in introducing nested layers of complexity and a steep learning curve.
A Closed Ecosystem in an Open World
While Kubernetes was born open, extensible, and cloud-agnostic, vCD was tightly tethered to vSphere + vCenter + vShield (later NSX) + vSAN. It never learned to breathe outside the VMware bubble.
A Failure to Inspire Developers
Unlike AWSP, Docker, or Kubernetes, vCD never inspired a developer movement. It was always ops-first, UI-driven, and bound to templates, not code, APIs, or declarative provisioning.
By the time Tanzu arrived, the damage was done. VMware had lost its chance to be the DevOps platform of choice. vCD had become the Lotus Notes of cloud infrastructure, technically powerful, but spiritually obsolete.
So, What Replaced vCloud Director?
Here’s what modern infrastructure looks like in 2024–2025:
-
Kubernetes everywhere, especially managed services like EKS, AKS, GKE, and OCI OKE
-
Infrastructure as Code as the norm: Terraform, CloudFormation, Pulumi, Crossplane
-
Multi-cloud orchestration with GitOps, ArgoCD, and Backstage
-
Serverless frameworks and event-driven architectures for modern SaaS
-
For remaining VMs: HashiCorp Nomad, Proxmox, or direct vSphere provisioning wrapped in Ansible
In other words: lightweight, composable, API-driven, and built for developer velocity.
The days of a centralized portal governing vApps with fenced networks and catalog templates? Gone.
VMware’s New Identity: Infrastructure Holding Company
As of 2025, VMware is no longer a software innovator. It's a portfolio brand. An enterprise infrastructure holding company under the Broadcom umbrella, offering:
-
Virtualization (ESXi)
-
SDDC packages (VCF)
-
Network & Security (NSX)
-
Storage (vSAN)
-
High-margin support contracts
It’s still profitable. Still entrenched. Still necessary for certain workloads.
But the soul of innovation? That torch has passed on to cloud-native players, to open-source maintainers, and to startups rethinking what infrastructure even means in the age of AI.
What We Keep and What We Let Go
If there’s one thing the rise and fall of vCloud Director teaches us, it’s this:
Abstractions must serve the user, not the vendor.
vCD tried to mimic AWS for VMware’s sake. Tanzu tried to retrofit DevOps into legacy. But in the end, the platforms that thrive are those that empower builders, embrace openness, and simplify the hard problems without hiding them behind wizard-based UIs and XML schemas.
I won’t say “I told you so.” (Okay, maybe a little.)
But I will say this: it’s time we stop designing tools for infrastructure tourists and start building platforms for explorers—ones who write YAML, not fill out forms. Ones who deploy with Git, not drop-down menus. Ones who build, not click.
Comments
Post a Comment