In a recent Gartner field search study, two early internal IaaS cloud adopters noted that if Amazon was the benchmark by which they are measured in terms of cost, then they had to make tough decisions regarding best-of-breed vs. good enough. In particular, the two clients cited whether deploying a third party virtual switch (i.e., Cisco Nexus 1000V) was absolutely necessary, especially if the cost made the internal cloud less competitive with Amazon. These organizations weren’t doing apple-to-oranges comparisons either. They came up with a per-VM cost broken down by both infrastructure and management/operations software. The cost of operational software was added to the Amazon cost to create an apples-to-apples comparison.
Enterprises are having to make tough choices regarding virtualization technology and all associated infrastructure and management products. To deliver cloud services, the enterprise has to be able to provide services quickly, securely, and reliably. In other words, the cloud service should come with the expectation “that it just works.” That’s a tall order for increasingly complex data center infrastructures. At this point, you may be wondering what any of this has to do with VMworld. Let me explain.
VMware made numerous data center and cloud related announcements at VMworld, including:
- New VMware vCloud® Offerings Accelerate Journey to the Enterprise Hybrid Cloud
- NetApp and VMware Help Transform IT, Make Cloud A Reality
- VMware Introduces New Platform for Enterprise Database as a Service
- VMware and Cisco Collaborate on Cloud Innovation
- Dell Announces Its First Public Cloud Offering; Dell Cloud With VMware vCloud® Datacenter Service
- Harris Corporation and VMware to Deliver Trusted Enterprise Cloud to Federal and Commercial Customers
I’m not here to dissect all of the announcements. For good perspectives on the vCloud Connector and Global Connect announcements, take a look at Lydia Leong’s and Kyle Hilgendorf’s posts. That being said, I wanted to comment on the body of work. VMware’s vCloud web site lists a growing number of provider partners, and many VMware customers I speak to about hybrid cloud state concerns about the need for hypervisor parity. That’s because they include the hypervisor as part of the application QA processes. As a result, they see it as less costly to move a VM between the same hypervisor type. I had blogged about this subject before. Bottom line – for many enterprises seeking mobility between data centers and cloud, VMware has a home court advantage. Other providers (e.g., Amazon) maintain the advantage for applications deployed straight to the cloud, with the enterprise having no intention to pull them back in.
VMware’s hybrid cloud strategy is quickly evolving, many customers are onboard with it, and at the same time, those customers are starting to question where they can save costs. Competitors such as Microsoft have their own thoughts on cost. Assuming organizations maintain a homogenous VMware IaaS cloud, that means that instead of trying to cut costs at the hypervisor/virtual infrastructure layer, they’ll look elsewhere. Again, if Amazon is the benchmark, the enterprise has to be sensitive to cost.
To VMware’s credit, they have been more transparent with partners regarding their strategic direction. There is no question that storage, networking, security, and management features that VMware considers essential to hybrid cloud infrastructure will be in the vSphere platform. I had lengthy discussions with two security vendors at the show, and they were comfortable in how they would innovate around vSphere moving forward. I got the same impression from the storage vendors I met with.
Today we have two general classes of cloud IaaS platforms: commodity “I don’t care” infrastructure, and enterprise “I do care” infrastructure. Enterprises use commodity infrastructure (e.g., AWS) for some workloads and enterprise (e.g., vSphere) for others (I know; it’s not that black-and-white. Stay with me). With an increasing number of features (VXLAN is the latest example) going into the hypervisor, one could say that VMware is creating a third tier – call it a “good enough enterprise tier,” or whatever you like. That tier, in my opinion, will try to compete with both the “I don’t care” and “I do care” infrastructure options. It will be lighter on third party value-adds and heavy on VMware products. This should concern some VMware technology partners. Their job is to convince customers that any “good enough” tier really isn’t good enough without their value-add.
If you’re a customer, you should be thrilled. Amazon has put down the gauntlet on cost, and the industry has to follow. VMware and other virtualization vendors (XenServer Intellicache is a good example) are commoditizing select infrastructure features that previously had come at a premium. This means that infrastructure software and hardware vendors have to step up their game. They have no choice but to innovate. At the same time, they have to be increasingly cognizant to the fact that “good enough” is becoming a more serious competitor.
At VMworld, VMware showcased a vision for a highly robust IaaS platform. We’re in a significant state of transition, and there will be some major winners and losers. If we go forward 10 years and VMware is the winner, then who are the losers? Or is VMware heading down the wrong path? I’d love to hear your thoughts.