As Cisco and Hewlett-Packard take the lead in defining what the next generation of the data center is going to look like, it's worth noting that the converged data center will be upon us sooner than most people realize.
Already, virtual servers are poised to outnumber physical servers by the end of this year. That means that the way IT organizations need to think about managing 'servers' and all the related IT infrastructure needs to fundamentally change. The simple fact is that the number of 'servers' that need to be managed will overwhelm the number of people available to manage the environment. And we can't keep adding people to the equation without driving the total cost of enterprise computing to unacceptable levels.
Cisco's answer to this challenge is to drive people towards upgrading to a new Unified Computing System (UCS) platform. HP is trying to work both ends of the equation with new Blade Matrix systems and a new Converged Infrastructure architecture that allows the concept of data center convergence to be applied to both existing investments and in new systems.
Other companies pushing a new platform approach to data center convergence include Intelicloud, while IBM, Dell and Sun are pushing their own versions of unified computing architectures that focus more on software to integrate server, storage and networking components, as opposed for calling for forklift upgrades.
In fact, the companies calling for forklift upgrades are relative newcomers to the server space. Free from having to support any legacy system architectures, companies such as Cisco and Intelicloud can be more aggressive about data center transformation. The incumbents, on the other hand, are not in a position to ask economically strapped customers to throw out every system investment they have made in favor of new server designs.
Driving all these changes are continuing advancements in virtualization. Not only are we increasing server utilization rates with virtualization, we're changing the way application workloads behave. With the advent of technologies such as V-Motion and V-App from VMware, application workloads will dynamically move across the enterprise. In fact, the Distrubuted Management Task Force (DMTF) is already working on an Open Virtualization Format (OVF) specification that should make the capabilities associated with V-Motion and V-App across all virtual machine implementations. So not only is the number of virtual servers growing, the complexity of managing the application workloads that run on them is about to dramatically increase in scope.
Regardless of how all this plays out exactly, radical change to the way we manage data center is coming. The complexity associated with managing server, storage and network resources is not only economically unviable, it hinders productivity because there is no single pane of glass through which all this equipment can be managed. As we all know, any change to one element of the data center has an immediate effect on the others, so managing them cohesively not only makes economic sense, it makes everybody's job easier.
The big challenge, however, will be getting there. Change is always threatening. But compared to the amount of cultural change the data center convergence represents within the IT department, the technology is going to be the relatively easy part.