Up In The Clouds : Helium not Hot Air

The greatest difference between the virtual data center and the cloud is how it is consumed. The greatest obstacle to the cloud can be a lack of synergy in how it is created.

Picture the house from the movie Up. In many ways I see the true innovation that will create the cloud and enable true infrastructure as a service (IaaS) as being a bit like this house. The effort to abstract infrastructure from physical and operational dependencies is like the effort of Carl, trying to free his house from its place attached to the earth. The breaking of this boundary is the demand for a better way to provide a service to business needs; to really provide infrastructure as a consumable service that truly aligns to financial and process models more effectively. Or, in a simple word: The Cloud.

The effort to break these dependencies has been from multiple sources. VMware created the robust hypervisor, breaking the physical boundaries to the server hardware and networking. VMware extended the ability of the hypervisor to enable the cloud by opening numerous APIs and facilitating integration across all layers.

Cisco extended VMware’s effort on networking by creating a virtual switch that added security, alignment of roles, and abstraction-aware application of policy. They have abstracted the identity of the servers themselves with UCS and changed the way x86 hardware is managed, provisioned, and designed. With DCB and FCoE Cisco is creating a physical network which can be a medium for multiple logical connections securely and with proper quality of service. And, Cisco is also extending networking as a logical component across datacenters with OTV breaking the logical addressing boundaries.

EMC is working hard on accelerating the hypervisor via leveraging VMware storage APIs and abstracting storage to allow data to be fluid and physically abstracted. With the implementation of storage-tiering using FAST, even the physical capabilities of the underlying disks will be abstracted from the workload. Furthermore the software stack to enable the consumption, operational change, security, compliance, and integration are available and high-priority goals within EMC.

All of these innovations are balloons attached to the house in an effort to break it free. Alone they will not do the job. It takes a focused drive by highly skilled people all working towards a single vision. And that is what is so important about the VCE initiative. When I look at my fellow teammates, VMware and Cisco included, I see individuals who are all focused on moving the ball forward (great Chad Sakac’ism). The vBlock is a great example of this effort. Single support, proven tested workloads, and tight integration from hardware through software is an important step in breaking free of the old paradigm.

A lot of the arguments and discussion on VCE can delve into battles over single balloons. A tactical, magnifying glass view over storage, servers, or hypervisor specific data points does not accomplish the ultimate goal. To really see the advantage you have to step back and look at the house. Storage, compute, networking, security, and operational management will constantly innovate over time individually. The push to a true cloud requires that these are focused and cohesive in order to innovate the way we all view infrastructure as a whole.

VMware, Cisco, and EMC are working diligently to lift the house. And I am just lucky enough to be one of the guys to whom they handed a tank of helium.

.nick