The second thing VMware needs to do is bring more clarity to a field that begs for leadership: virtual networking. VMware in effect has opted for a collaborative, group effort in virtual networking. It bought Nicira for $1 billion, then encouraged Martin Casado and other Nicira talent to continue contributing to the virtual networking project of OpenStack. How will VMware capitalize on the rapid buildout of a virtual networking platform being constructed there? It's now clear that virtual networking will fit into the larger virtualized data center, but it's not clear where customers should first invest to bring that resource closer to reality. Rapid gains were realized in the consolidation of x86 servers on VMware hosts; can those gains be repeated in networking?
Third, VMware is too intensely focused on the role it can carve out for itself in cloud computing. It has a natural role to play as the purveyor of the private cloud, which requires extending the virtualized data center into a more standardized, policy-driven operation with end-user self-provisioning. The need for the company to play a larger role in the public cloud is less clear, even though it's succeeded in convincing everyone that it feels it must at all costs be a major player there as well.
I'm not so sure. VMware with its grasp of virtualization is a natural data center systems company, and it will be able to extend its position there into hybrid cloud computing in several different ways. One would be to let Amazon Web Services -- and eventually other cloud vendors -- accept VMware workloads, which Amazon already does. Allowing this to set off a burning sense of competition with Amazon puts management in the position of saying that it's not only the best at virtualization; it will be the best at public cloud computing too. And that puts more management credibility on the line than is perhaps warranted, given the head start and capabilities of the competition.
VMware will pick up its share of the hybrid cloud market. Some customers will want an all-VMware world and will want the assurance of transferring their workloads from one vCloud Director to another as they take advantage of the public cloud. But there are going to be other solutions, and getting this issue into perspective would go a long way toward convincing customers that VMware has got its priorities straight. Most of them didn't know VMware was worried sick about cloud competitors until Gelsinger and Carl Eschenbach told them so at the Partner Exchange in April.
Fourth, accept the reality that it's a multi-hypervisor world and there's nothing wrong with that. Windows and Linux rule the data center, which means Microsoft will slowly but relentlessly take market share. The future of virtualization is managing all those virtualized resources as a single operation, including Hyper-V. Gartner's Magic Quadrant says Oracle VM is edging up as a challenger, but I believe the bigger threat is KVM, still down among the niche players. VMware should do everything it can to be the level-headed broker of these resources as well, and concentrate those management capabilities behind its concept of the software-defined data center.
Those who assume virtualization is a saturated market, and that it's inevitable it will move to commodity products, don't understand how much virtualization has changed things, how much it will continue to change things and how much some customers will continue to rely on VMware to lead those changes far into the future.