Posts Tagged Marcia Kaufman
Many business executives are interested in moving to the cloud because of the potential impact on business strategy. Increasingly they are convinced that a cloud model – particularly the private cloud – will give them an increased amount of flexibility to change and manage the uncontrolled expansion of IT. In contrast, from an IT perspective, the ability to virtualize servers, storage, and I/O, is often viewed as the culmination of the cloud journey. Of course, the world is always more complicated than it appears. Cloud computing if implemented in a strategic manner can help a company experiment and change more easily. Likewise, virtualization, which may seem like an isolated and pragmatic approach, needs to be considered in context with an overall cloud computing strategy.
However, the challenge for many companies right now is how to transform their virtualized infrastructure into a private cloud that delivers on the promise of on-demand and self-service provisioning of IT resources. To make sure that business leaders gain desired cost savings and business flexibility while IT gains the optimization that can be achieved through virtualization requires an integrated strategy. At the heart of this strategy is service management of this emerging highly virtualized environment.
Why is service management important? There are good reasons why a key focus of the virtualization strategy at many companies has been on server virtualization. For example, server virtualization helps companies create a faster and more efficient IT provisioning process for users. It helps users with increased operational flexibility based on the mobility and isolation capabilities of virtual machines.
However, there is a major management problem with the typical virtualization approach in many companies. Often developers satisfy their demands for computing resources by simply creating or spinning up a new virtual machine rather than anticipating that they might not have the time or money to purchase new IT systems. IT management in the beginning allows this practice because it is easier than trying to control impatient developers. However, there is a price to be paid. Each new virtual machine image requires memory and disk resources. When the number of virtual machines grows out of control, companies end up spending more time and money on disk, storage, and memory resources than was anticipated. Lack of control means lack of management. What is the answer?
At a recent IBM systems software meeting, Helene Armitage, General Manager of Systems Software, emphasized that in order to truly transform the data center into an agent of change requires a focus on manageability. Here are the main take always from my conversations with Helene and her team.
- Virtualization is not just about server consolidation. Companies need to think differently about virtualization in order to prepare for a future marked by rapid change. For example, storage virtualization can offer big benefits to companies facing explosive demand for data storage from increased use of new technologies like embedded and mobile devices.
- Physical and virtual environments will need to be managed in a unified way and at the same time. Some may think of virtualization as a way of reducing management requirements. After all, with consolidated resources you should have less demand for power, energy, and space. However, management needs to be done right to be effective. it is important to consider how to manage across the platform and to include monitoring and configuration in combination with business service management – delivering IT services in a structured, governed, and optimized way.
- Cloud changes the complexion of what’s important in how virtualization is approached. In the cloud, virtualization takes on a different level of complexity. If elements of virtualization including storage, server, and I/O virtualization are handled as isolated tasks, the cloud platform will not be effectively managed.
- System pooling is key for managing lots of systems. Systems pooling essentially means that IT treats all resources as a unified set of shared resources to improve performance and manageability.
- Enforcing standardization right from the beginning is a requirement — especially when virtualization is at the core of a cloud environment. The only way to truly manage the virtualized environment is to apply repeatable, predictable, and standardized best practices across the entire computing environment.
The greatest challenge for companies is to think differently about virtualization. Management will need to realize that virtualization is a foundational element in building a private cloud environment and therefore has to be looked at from a holistic perspective.