VMware's Hybrid Cloud Service, detailed Tuesday, has important differences compared to Amazon Web Services. Customers will have to weigh the partner model and pricing.
VMware will establish four data centers from which it will offer its competition to Amazon Web Services (AWS) infrastructure-as-a-service, the company announced Tuesday. But there are big differences between what VMware is doing and what an AWS, Google or Rackspace does.
Not one existing Amazon customer is likely to abandon EC2 to take up residence in a soon-to-be-launched VMware facility. But many existing VMware customers are likely to remain comfortably in the VMware fold because of those four centers, if they work out as planned.
VMware wants partners to supply compatible public cloud facilities, but has been frustrated at the lack of sophistication and willingness to adopt full-scale VMware implementations. Bill Fathers, VMware's new senior VP and general manager of its Hybrid Cloud Services unit, acknowledged in an interview he was hired to build up the capabilities of the VMware ecosystem, and that includes training and equipping more hybrid cloud providers. He was recruited for the task in March from a seasoned cloud supplier, Savvis, now a unit of CenturyLink.
While it looks like VMware wants to take hybrid cloud business away from partners, it literally can't provide an Amazon-scale alternative to the public cloud and must rely on partners to fill in many gaps. That makes establishing four VMware data centers something of a balancing act.
VMware is also in a poor position to compete by building ultra-modern data centers, as Facebook did in Prineville, Ore., and Forest City, N.C., and then offering low-cost compute cycles out of such infrastructure. On the contrary, VMware won't build anything. It will lease space from wholesale data center builders. It will then wheel in racks of servers, most likely from its Virtual Computing Environment (VCE) subsidiary, based on partner Cisco's converged compute and networking infrastructure, and throw on the switch.
One such data center will be located in Santa Clara, Calif., a leading site for wholesale space built by CoreSite, Digital Realty Trust and Vantage. Other VMware centers will take shape in Las Vegas, Dallas, and Sterling, Va. These facilities, to some extent, represent the opposite of those that Facebook, Google and Amazon have built. Those three companies compete from their own supersized facilities, stacked with servers built to their own designs. VMware will rely on smaller leased facilities using more standard parts.
Renting third-party space, VMware will be able to establish such centers quickly. Left unsaid is that they can phase-out leases and wheel away equipment almost as fast as they wheeled them in, if, for example, demand slackens or customers express more interest in the public cloud elsewhere. But for now, VMware isn't going anywhere but straight into hybrid cloud computing.
One possible outcome is that these facilities will serve as a model for VMware partners to emulate, and within three to five years the partners will do such a good job that VMware won't want to operate its own hybrid cloud service data centers in perpetuity. VMware CEO Pat Gelsinger, given the chance to comment further after the announcement, said he saw the hybrid data centers as both a role model and a place where VMware "could set the bar higher and higher" for implementing hybrid services, based on its evolving product set. For now, VMware believes it can implement a fuller set of services and make them easier to use than partners can. And it wishes to do so to forestall more customer movement to AWS.
Fathers amplified that description further in an interview after the announcement: "By becoming a service provider ourselves, we will learn a lot of important lessons that we can bake into [vCloud Director]."
Multicloud Infrastructure & Application ManagementEnterprise cloud adoption has evolved to the point where hybrid public/private cloud designs and use of multiple providers is common. Who among us has mastered provisioning resources in different clouds; allocating the right resources to each application; assigning applications to the "best" cloud provider based on performance or reliability requirements.