X

How a private cloud goes beyond virtualization management

Both in terms of their underlying approach to IT and their management capabilities in the aggregate, there's a clear distinction between virtualization and private clouds.

Gordon Haff
Gordon Haff is Red Hat's cloud evangelist although the opinions expressed here are strictly his own. He's focused on enterprise IT, especially cloud computing. However, Gordon writes about a wide range of topics whether they relate to the way too many hours he spends traveling or his longtime interest in photography.
Gordon Haff
3 min read

At first blush, private clouds can look a lot like virtualization. But first looks can be deceiving. While it's certainly true that we see virtualization management products extended--for example, with self-service portals--in ways that make them look superficially like clouds, we're really talking about different categories of software.

There are two ways to think about these differences. The first is in terms of different mindsets and approaches to IT operations. The other is to consider specific features and capabilities.

Consider the "big picture" aspect first. Mary Johnston Turner, research vice president of enterprise system management software, at market researcher IDC contrasted virtualization and cloud at the IDC Directions conference last year. She described virtualization as providing the "underpinning for cloud" while she described cloud as going "beyond virtualization to focus on services and consumption."

Expected adoption of cloud management building blocks. IDC

What does this mean exactly? In their book Visible Ops: Private Cloud, Andi Mann, Kurt Milne, and Jeanne Morain write that it's a "shift from framing virtualization expansion goals based on footprint and consolidation metrics to offering business-optimized services. Services that meet user needs will drive adoption. Widespread adoption is critical to driving cloud economics."

This is done by designing a catalog of standardized services--think of them as application or development environments--and offering them to consumers, such as developers, through a low-touch self-service interface. Access to these services is controlled by policy, as is the runtime management (such as patching) of these environments after they are deployed.

Virtualization management, and the ecosystem of third-party management products and add-ons that have grown up around virtualization over time, certainly bring a degree of order and process to virtualized environments. Their goal is to reduce what is sometimes called "VM sprawl" resulting from proliferating unmanaged virtual machines. However, I'd argue that policy, lifecycle management, and standardized workflows are more fundamentally baked into cloud solutions.

Furthermore, cloud solutions that operate independently of the underlying virtualization layer can span hypervisors and other platforms such as public clouds. This allows management and policy to be extended over a more heterogeneous set of resources.

At IDC Directions, Turner also showed data on the adoption of "critical cloud management building blocks" that reflected "the percentage of IT decision makers reporting their organization is/plans widespread or selective production use of all technology."

Unsurprisingly, virtualization management was the most common current or planned technology. It was present to some degree in about 70 percent of cases with modest (a few percent) growth in penetration forecast by 2012. Automation/orchestration and self-service portals were the next most common building blocks, growing from about 40 percent of organizations to over 50 percent. Service management/service catalogs came next; end-to-end performance/availability management and consumption metering came last in terms of adoption with only about 30 percent anticipating metering even by 2013.

This data is consistent with my observations. A certain amount of automation and self-service can be put in place to augment virtualization but it's with service management and catalogs that IT operations really begin looking like a cloud provider to their internal consumers. This is the point where we stop thinking so much in terms of managing servers and more in terms of operating IT-as-a-service.

It's also been my observation that public cloud-like charge-back and "showback," while of considerable interest to some organizations, potentially represents a big change in the way that the Lines-of-Business pay for IT. Organizational changes like this happen rather quickly and it may never fit with the way some companies want to do internal accounting. Therefore, I don't find it surprising that adoption rate is likely to trail.

Virtualization management solutions will continue to add features over time. And these will often overlap with what private clouds do on an individual feature basis. However, looking at both the underlying approach to IT in the big picture and management capabilities in the aggregate draw a pretty clear distinction between virtualization and cloud.