I had the pleasure of attending VMworld this year in Las Vegas and the overall message was clear: hybrid cloud needs to be a real, usable, manageable, sustainable thing, and VMware wants to make it work. It's both interesting and validating that Nutanix had the same message at .NEXT a couple of months ago in Washington D.C. Whereas Nutanix is building an Acropolis-based cloud in the Google (GCP) datacenters, VMware has hitched their wagon to Amazon Web Services.
Before now, customers have been deluged by the "bimodal" messaging: bursting, fast changing, customer-facing workloads in the public cloud (usually AWS, but also Azure and GCP) and sustaining, slow changing backend workloads in the private cloud (or plain old server virtualization). That's sort of being done, and I have several customers who maintain separate organizations within their enterprise to support and maintain their disparate public and private workloads. But that is not what most of us want when we talk about "hybrid cloud." It doesn't deliver on the promise of actually being able to move an existing workload or spin up a new workload in either location. Until now, we would end up using different tools to provision on-prem vs IaaS, and we'd have separate, complex networks that looked different depending on where the workload was running. Along with the myriad of options out there for running a workload in the public cloud, many customers (particularly those who still run their workloads in their datacenters) get stuck figuring out how to start moving into the public cloud.
Now we're seeing VMware and Nutanix pick their horses and start riding. VMware is essentially saying, "If you're a VMware shop (and who isn't?) and you're considering moving workloads to a public cloud service (and who isn't?) then you can turn up VMware Cloud on AWS (also known as “VMC”) and extend your enterprise into AWS while using the same tools and the same skills." Your VMware admins will recognize the management interfaces. Your VM's will be on .vmdk's and connected to NSX-based logical networks. You'll still call VMware for support. And the big differentiator between this and IaaS from a colo provider is that native AWS services can be accessed seamlessly.
Nutanix presented a similar message at .NEXT. They're positioning themselves as the Apple to VMware's Android. VMware is everywhere, and more flexible, and more complicated. Nutanix is simple and more opinionated. If you're a Nutanix customer, you can click a button and extend your Nutanix system into their DRaaS offering (called “Xi”) run in GCP's datacenters. It is presented to us as more tightly a Nutanix service than VMware, which is presented rather as more of a partnership with AWS. There are differences, not the least of which is the scale of the customer base, but they underscore each other's messages: hybrid cloud needs to work, and needs to be easier.
Another piece of this message from VMware was in Pat Gelsinger's keynote when he said that NSX would be to VMware for the next ten years what vSphere / ESX was for the previous twenty. This admission was a pretty big deal. I think we've seen the industry really get to "good enough" in other hypervisors and we've watched VMware try to determine what new features in vSphere will keep customers renewing and expanding their vSphere deployments. We've also seen NSX become the leader in customer SDN, allowing us to accomplish network automation, microsegmentation, and extend our logical networks across datacenters, and now up into AWS. This will be a key piece of making hybrid cloud or multi-cloud workable.
Overall, this year is quite exciting. It feels like the industry is slowly crossing a tipping point where we'll see more legacy workloads move into public cloud services, or at the very least connect more cleanly to newer workloads which live there.