It has been more than two years now since Broadcom spent $69
billion to buy virtualization giant VMware and in short order rattled the cages
of enterprises and the wider tech industry by announcing significant price
increases, new licensing plans, and a host of other changes. The company had
a plan for what they wanted to do with VMware and has moved forward, with
the result being anger, shock, and recriminations.
By some accounts, the result has been less the expected
stampede of customers out the doors and more reassessment of their virtual
machine (VM) strategies, what executives with one hybrid cloud management
platform vendor, CloudBolt, called a “slow unwind” after the initial reaction.
CloudBolt’s survey earlier this year found that 86
percent are reducing their VMware footprint and 88 percent are concerned
about future price hikes. Fifty-four percent said they’re staying with VMware
reducing their dependence. About 72% of workloads being migrated are
moving to the public cloud.
As one survey respondent said, it’s taking as long as two
years to unwind a decade of process dependencies.
Coming To The Rescue
Over the past two-plus years, datacenter IT vendors have
been looking for ways to attract these discontented companies as they configure
their strategies. Firms like Red
Hat and HPE are offering VMware customers soft landing spots.
HPE in 2024 unveiled HPE VM Essential, a package of
virtualization capabilities that included a hypervisor and VM management, and
last year combined
VM Essential with its Morpheus cloud and IT management software, folding
them both into its HPE Private Cloud Business Edition in a move that the
company said would reduce the cost of VM licenses 10 times.
This week, HPE took another step with the fourth generation
of its Private Cloud platform that moves it closer to being a single, unified
operating model for running modern workloads – including AI-based jobs –
stretching from core datacenters to the cloud and out to the edge with form
factors from hyperconverged to disaggregated.
“What this means is that you’re going to have a single pane
of glass, you’re going to have one place where all of these organization actors
can interact and do their workloads,” Angel Pinella, vice president of private
cloud product and engineering at HPE, told journalists. “You’re going to those
that are consuming VMs and they’re consuming … by self-service or by those
containers, as well that they’re going to be needing for their development
operations. You’re going to have those that are setting policy, you’re going to
have those setting the security as well. There are a lot of hands in the pot
and if you don’t have one operating model … it just doesn’t work. It’s going
to take a long time and it’s going be very fragmented.”
Unifying Container And VM Management
The Private Cloud platform, built atop HPE’s
ProLiant Compute Gen12 servers announced last year, unifies the management
of VMs and Kubernetes containers via a single interface in Morpheus, and
supported VM technology not only includes VM Essentials but also VMware virtual
machines, throwing a line to those in CloudBolt’s survey that are reducing
their VMware footprint or staying with VMware but reducing their dependence, or
catching some of the workloads migrating to the cloud.
The unified management of containers and VMs is due to be
generally available in the third quarter. Having a single management model for
VMs, containers, and AI offers an antidote to what Pinella said has been a
disjointed collecting of various management tools across those environments
that has led to unneeded complexity and costs. VMs, containers, and AI
workloads need to be first-class citizens in IT environments, which requires a
single management layer combined with a choice of runtime and platforms.
“HPE is different because we’re offering choice in
hypervisor,” he said. “There were the days where one hypervisor was totally
fine. Those days are over. You have unique, different hypervisor footprints for
different footprints physically in different places in the network, whether
it’s in a co-lo or whether it is at the edge, which is one of the most emerging
topics that we talk to CIOs about today. For that distributed edge, that
distributed enterprise, a lot of the data that’s being consumed, manipulated,
and inferenced on is being created at the edge. A lot the virtualization
platforms out there just don’t cut it and so we want to be as nimble as
possible to be able to provide the right level of virtualization footprint for
the right level of location and network and all of this needs to be delivered
with built-in resiliency security data protection and that’s where HPE
differentiates.”
HPE isn’t the only one looking to Kubernetes as a way to
reduce cost and complexity in cloud infrastructures and the growing prominence
of AI and other advanced workloads. Broadcom executives early last month
boasted of the latest
version of VMware’s vSphere Kubernetes Service (VKS) Kubernetes runtime as
a tool to address such problems.
Taking A Run At VMware Users
That said, HPE – like others – sees an opportunity to entice
organizations to make the move away from VMware. In addition to the unified
management of containers and MVs, HPE’s Zerto software, which offers data
protection and disaster recovery capabilities across hybrid cloud environments,
now offers organization live workload migration from VMware environments to
HPE’s VMs with little disruption, another avenue for VMware users looking for
an off-ramp.
HPE also developed a streamlined upgrade path to the
Morpheus’ enterprise edition and said SimpliVity, its hyperconverged
infrastructure platform, now supports Morpheus VM Essentials.
Beyond the Private Cloud enhancements, HPE also unveiled
improvements to its Alletra storage line, with the Alletra Storage MP X10000
providing new native file storage along with the object storage already in the
platform. The system can scale to 16 nodes with 23 Petabytes of capacity and deliver
remote direct memory access (RDMA) file storage, an addition to the existing
support for S3 object storage over RDMA.
