Cloud Computing = Managed Services, Measured In A Different WayCloud Computing = Managed Services, Measured In A Different Way

Just what IS cloud computing, anyway? Is it really all that different from managed hosting, ASPs, client-server, and all the rest? According to a managed services provider I spoke to recently, the difference increasingly lies in how it's measured and sold.Simon West is Chief Marketing Officer of Terremark, a Miami-based managed services provider with a focus on the mid-market. His company has recently introduced The Enterprise Cloud, which offers cloud-based IT infrastructure to midsize companies. The service lets companies quickly provision processing, storage, and networking resources on demand via a Web-based interface.

What I wanted to know, though, is how that really differs from Terremark's original business of traditional managed services. Is it just a name change? According to West, the answer is yes and no. As virtualization takes hold in data centers large and small, the units of measurement when buying these services are changing. Instead of buying by the the fairly large unit of a physical server, storage array or network pipe, for example, you're buying in the far more granular units of GHz of processing power or bandwidth and gigabytes of storage. With virtualization, those resources are located on dynamically changing bits and pieces of the entire data center -- or across multiple data centers. You don't know, moment to moment, which physical devices you're actually using. The difference is largely transparent to end users, but according to West it's the key difference between managed services and cloud computing infrastructure. The advantage, of course, is that the "cloud" approach makes it much easier to scale resources to meet spikes of demand. West said that using the Web interface, Terremark's customers can provision new resources in minutes, not days or weeks. It's also much more efficient and more reliable. West estimated than corporate server CPU utilization tops out at around 8% -12%, for example, but can go much higher in a shared, virtualized datacenter. And when the resources are spread around, there's no single point of failure. More From bMighty: Running Your IT Infrastructure In The Cloud