Broadcom’s decisions to replace perpetual VMware software licenses with subscriptions and to eliminate point products in favor of an expensive bundle of private cloud tools are driving longtime VMware customers to look for an exit strategy.
Many clients have reported dissatisfaction with the new licenses, according to research firm Gartner, which notes that changes have forced companies to purchase bundled VMware software they don’t intend to use. “For the typical customer, Broadcom’s ‘modernization’ of VMware license pricing has increased VMware run costs by a factor of two or more,” writes Gartner analyst Paul Delory in a report about VMware replacement options.
For customers looking elsewhere, there’s no shortage of alternatives. Delory has identified six potential offramps: Hyperconverged infrastructure (HCI), public cloud infrastructure as a service (IaaS), stand-alone hypervisors, distributed cloud, container orchestration, and open infrastructure.
The problem is finding alternatives that provide a commensurate level of performance, reliability, security, management and risk reduction. “There is no like-for-like replacement for the VMware hypervisor on the market,” Delory points out.
IDC analyst Stephen Elliot concurs. “VMware works and works really well. It’s enterprise-grade. It’s scalable. They’re still the Lamborghini.”
This dilemma of whether to absorb the Broadcom price hikes or embark on the arduous and risky journey of untangling from the VMware ecosystem is triggering a broader C-level conversation around virtualization strategy.
“For enterprises navigating this uncertainly, the challenge isn’t just finding a replacement for VMware. IT shops of all sizes see Broadcom’s actions as an opportunity to rethink their approach to virtualization, cloud strategy and IT modernization,” says Steve McDowell, chief analyst at NAND Research.
Elliot says that server virtualization has been taken for granted for a long time, and the Broadcom-driven wake-up call is forcing organizations to reevaluate their virtualization strategies at the board level. “That kind of strategic conversation hasn’t happened for years. Customers are saying, ‘What can we do as this platform emerges from VMware. How do we think about this relative to our multi-cloud strategy and private cloud and the efficiencies we can gain? Let’s talk about risk reduction. Let’s talk about platform strategy.’ This is an opportunity to identify business value. It’s triggering this plethora of swim lanes.”
Check the waters before diving in
While there are multiple alternatives to the VMware platform, none of them are as good from a feature perspective, and there’s a risk associated with moving off a tried-and-true platform.
In estimating the cost of a large-scale VMware migration, Gartner cautions: “VMware’s server virtualization platform has become the point of integration for its customers across server, storage and network infrastructure in the data center. Equally, it is a focus of IT operational duties including workload provisioning, backup and disaster recovery. Migrating from VMware’s server virtualization platform would require untangling many aspects of these investments.”
It would take a midsize enterprise at least two years to untangle much of its dependency upon VMware, and it could take a large enterprise up to four years. Even then, the cost associated with planning and executing the migration might eat up any savings associated with lower licensing costs and might introduce additional risk, says Gartner.
McDowell agrees. “Moving away from VMware is no simple task. Enterprises must weigh migration complexity, integration challenges, and the long-term viability of their chosen alternative. The decision isn’t just about cost savings – it’s about aligning IT strategy with the future of hybrid cloud, containerization and AI-driven workloads.”
Forrester analyst Naveen Chhabra has developed an approach for thinking about server virtualization strategy across two dimensions: “time to license renewal” against “number of VMware software tools in use.”
Chhabra points out that the VMware bundle called Virtual Cloud Foundation (VCF) includes products in 12 different categories, including vSphere for compute, vSAN storage, NSX networking, plus management, automation, a Kubernetes service, virtual firewall, log management, cloud cost management, etc.
If a customer is only using one or two of those products and their license doesn’t expire for several more years, then there’s time to plan for, pilot and deploy an alternative. Conversely, if a customer is already using seven or eight of those products and their license expires within the next year or two, jumping ship would probably not make much sense, or even be feasible from a technical perspective.
“You might want to get off the VMware highway,” says Chhabra, “but don’t get attracted by the novelty and promises of the next technology. Try them out, do pilots. Make sure the platform serves business requirements. Let this not be a knee-jerk reaction to price increases.”
What are the alternatives?
Here are the alternative routes that enterprises could take, each with their own pros and cons. Delory frames the argument this way: “Since no competing hypervisor on the market is at feature parity with VMware, the question is whether the possible alternatives are good enough at a lower price.”
1. Replace on-prem VMs with public cloud infrastructure
There’s an argument to be made for a strategy that reduces reliance on virtualized on-prem servers altogether by migrating applications to the public cloud.
Delory points out that “building and migrating to any VMware alternative will involve some level of re-engineering and a considerable investment of money, time and effort. Those resources are probably better spent re-architecting applications to remove the need for virtual machines (VMs). Lift-and-shift to public cloud infrastructure as a service (IaaS) has become a much more attractive alternative.”
“Ideally, organizations moving VMware VMs to the public cloud would do so as part of an overall cloud transformation program, where the applications are modernized and VMs rendered unnecessary,” Delory adds.
Of course, most organizations typically have a stubborn group of applications that can’t be moved to the cloud for a variety of security, regulatory, compliance or other reasons.
2. Hyperconverged infrastructure (HCI)
HCI provides the full stack of software-defined infrastructure: virtual compute, storage and networking, plus infrastructure automation, monitoring and cloud management capabilities. HCI vendors include Nutanix, Scale, Microsoft Azure Stack and others.
Since HCI products are the closest equivalents to the VMware stack, they can be deployed with less effort than other solutions in terms of workload re-architecture and staff retraining.
However, switching to HCI is a capital expense decision. It requires buying new hardware, which could end up negating any cost savings associated with getting off the VMware bundle. Also, HCI products generally do not support external storage, and often limit the hardware on which the HCI software can run.
Although no HCI product can match the depth and breadth of VMware’s technical capabilities, they generally satisfy enterprise-level requirements or can be augmented with third-party tools until they do. For organizations that are reluctant to adopt a full-stack HCI infrastructure, the next option would be simply swapping out hypervisors.
3. Standalone hypervisor
There are a number of standalone hypervisors on the market, including Microsoft Hyper-V, Oracle Linux Virtualization Manager, and open-source KVM.
Standalone hypervisors are an option for customers only looking to replace ESXi and vCenter. But, again, standalone hypervisors can’t match VMware, particularly for storage management capabilities.
Gartner points out that vSphere Storage APIs Array Integration (VAAI) offers robust standardization and control of external storage, and vSphere Virtual Volumes (vVols) allow the hypervisor to manage a VM’s virtual disks directly. Other hypervisors largely rely on the array to manage storage independently of the hypervisor, which is riskier, more error-prone, and requires more time and effort to manage.
4. Container orchestration
Another approach is to deploy a Kubernetes container orchestration tool to manage VMs.
The open-source KubeVirt project wraps each VM in a lightweight container. Kubernetes then schedules the container, with the VM inside, enabling a single management toolchain for containers and VMs.
Gartner points out that KubeVirt is “new and unproven” with no track record of enterprise-grade performance and scalability. So, while it might be too early to take a chance on KubeVirt, stay tuned.
“KubeVirt is the basis of major infrastructure offerings from prominent vendors, such as Google Anthos and Red Hat OpenShift Virtualization. These extend KubeVirt with new capabilities and mitigate its risks with enterprise-ready support. With major vendors investing money, time and effort into the KubeVirt project, it is likely to grow and mature rapidly. In the next two to three years — which is many VMware customers’ target time frame for a replacement — KubeVirt may become a viable VMware alternative,” says Gartner.
5. Distributed cloud or managed private cloud
Another alternative is to bring native cloud services on-prem. Examples include AWS Outposts, Google Distributed Cloud Edge, or Oracle Distributed Cloud. With distributed cloud, public cloud vendors deliver their services to customers’ data centers, while the operation, governance and evolution of these services remain the cloud provider’s responsibility.
Enterprise customers would take this route when they have apps that need to remain on-prem for regulatory, security or latency reasons and don’t want to build their own private cloud. The downside is that the vendors “rarely makes the full set of public cloud services available on-premises,” so the adoption rate thus far has been low, according to Gartner.
6. Open infrastructure
OpenStack, the open-source virtualization infrastructure platform for both public and private cloud deployments, is a less viable option for the average enterprise because it has proven to be complex to manage. Yes, it’s free, but it also requires hiring OpenStack experts to operate the virtual infrastructure. Gartner says that despite some initial promise, OpenStack is deployed primarily in niche use cases, mainly research computing, hosting and telecommunications.
7. Stay the course
After enterprises run through all of the options, one outcome might be a decision to stick with VMware as the cornerstone of a broader private cloud strategy.
Elliot turns the idea that organizations shouldn’t pay for a bundle of tools they don’t use on its head. He argues that if an enterprise is already paying for these features, why not use them to build a private cloud that delivers public-cloud levels of efficiency, automation and business value, he says.
Elliot adds: “There’s no harm in saying to Broadcom, ‘We’ll do a 1, 2, 3-year renewal. Show us what the roadmap looks like. Prove it to us.’ This is an opportunity to reengage with VMware. If it works out, customers can say, ‘We delivered a lot of success here.”’
VMware had historically presented customers with a lot of complexity between its many product lines. By bundling those products into one package, Broadcom has simplified its portfolio, Elliot adds. “That’s a big deal,” says Elliot. And while there have been several high-profile defections from Broadcom, Elliot also points out that 30 million cores have been renewed.
Enterprises have more options that ever before, says NAND Research’s McDowell. “Organizations must carefully evaluate their next steps, whether moving to public cloud, adopting a hybrid model, transitioning to a new hypervisor or staying with VMware.”
“The time at act is now,” McDowell says. “Enterprises that proactively navigate this transition will mitigate the uncertainties of VMware’s new ownership and position themselves for long-term success.”
Source:: Network World