Pat Gelsinger stunned the enterprise computing world recently with the announcement of his jump from VMware to Intel. Pat will no longer be the man leading VMware – a technology leader and innovator that has, during his tenure as CEO – built a tenacious hold on enterprise computing. During his twelve years at VMware, the company has to $10.81B in revenue and established itself as a dominant player in enterprise IT.
It is important to understand this move within the context of enterprise computing today. While digital transformation has become a cliché, enterprise IT organizations are nevertheless seriously engaged in digital transformation projects. And for various reasons, COVID-19 accelerated their progress during the past year. VMware executives have made digital transformation a primary focus to ensure that at least two of VMware foundational virtualization technologies – ESXi and vSphere – remain firmly ensconced within enterprise data centers. But, as we enter 2021, the challenges to VMware’s hold on enterprise computing have never been greater.
VMware’s core technology has been and remains virtualization. Since the introduction of vSphere in 2009, the company has withstood challenges from Microsoft (Hyper-V) and the open source community (KVM and OpenStack) surprisingly well. But now, enterprise IT is increasing looking to use containers as their core virtualization technology. Among the motivations to transition from VMware’s virtual machines to containers is the relatively high cost of a VMware operational environment. Kubernetes has come on strong as a lower cost open source contender. Now with the backing of Red Hat/IBM and a list of other IT luminaries, enterprise-grade Kubernetes is certain to work its way into prime time operational environments. But as the competition heats up, Gelsinger exits. Can the next CEO guide VMware through this competitive battlefield?
When VMware appeared, it was common for operations staff to run dedicated servers to a single application or workload. This led to a situation commonly referred to as server sprawl which limited the overall scalability of IT service delivery and steadily escalated the cost of running a data center. VMware’s server virtualization platform enabled operations staff to consolidate multiple applications to fewer servers, allowing them to scale services delivery while better managing cost. At VMware customer events, users would brag about how much of their application portfolio they had virtualized.
Public clouds are able to use infrastructure virtualization to reach enormous scale. Unfortunately for VMware, they roll their own virtualization software. Yes, VMware users can run VMware operational environments in the cloud, but doing so is typically seen as an interim step to running workloads natively in the cloud. Once again, VMware faces shrinking opportunities in the enterprise data center as users migrate their workloads – many from servers running VMware software – to the cloud.
IBM built the first virtualization engine. It was called VM and it ran on a mainframe. Later iterations (VM/VSE and MVS) expanded IBM’s dominate position in within the data enter. But users came to chafe at the considerable expense of IBM’s virtualization software and rebelled. They directed new workloads to UNIX servers client/server architectures. Mainframe hegemony was at an end.
Is VMware now walking the same path?
The high price of VMware based computing fueled the rise of OpenStack. And while OpenStack failed to deliver an operational environment that the enterprise really wanted to use in production, its emergence signaled user’s discomfort in paying the VMware “tax.” Container-based virtualization managed by Kubernetes is now the more viable alternative and the majority of enterprises will likely have it in some stage of production by the end of this year.
Containers and Kubernetes is to VMware what client server computing was to the mainframe. However, to the credit of VMware executives led by Gelsinger, a defensive strategy was formulated that leveraged VMware’s strength in data center operations. Containers were coming to displace VMware VMs. But if IT operations staff was given a way to manage containerized workloads in a way that was familiar to them, VMware could essentially surround containers in a way that allowed the two to coexist and prosper. This strategy is now embodied in VMware’s Project Tanzu and its adoption is accelerating. Which is why Gelsinger’s exit and two other senior executives from VMware at this critical juncture is concerning.
In the years that Gelsinger has been in command, VMware’s performance has been nothing short of spectacular. He is without doubt, the executive most capable of guiding the company through the challenges presented by the rise of Kubernetes and cloud computing. His departure may be good news for Intel, but it is not a fortuitous event for VMware.