Jump Start with Virtualization Services

For many large companies, the decision to virtualize their X86-class servers is no longer a matter of if, but rather, of when. Many of IBM’s enterprise customers began the arduous task of determining if virtualization—using VMware® on Intel® and AMD-based platforms—was reaching the stage of being production-ready some 18 months ago. That’s when their IT departments began running their proof-of-concept tests.

For many, virtualization itself is a well-recognized and regarded technology for IBM System p™ and System z™ servers. What is less talked about, is the wide-spread use of virtualization on X86-class servers. The high acceptance rate by these IT groups of the technology is a big endorsement. More importantly, customers understand that maximum business benefit comes from more than virtualization alone: The bells and whistles of a complete solution are needed to make a rock-solid argument for virtualization. Simply “moving the mess for less” is not an option for clients cognizant of what benefits virtualization can bring. IBM not only offers the complete solution that businesses need to implement a full virtualization strategy, but also, through IBM Global Services, they provide a full range of assistance that can get the job started or the project completed.

"IBM not only offers the complete solution that businesses need to implement a full virtualization strategy, but also, through IBM Global Services, they provide a full range of assistance that can get the job started or the project completed."
Virtual Discovery
Getting the most out of virtualization requires serious upfront planning. What’s more, when that planning is done correctly, the results can put IT on the fast track for acting as a service provider. That’s the reason why many large customers bring in IBM Global Services to help when it is time to implement virtualization on X86-class servers.

For these customers, the pivotal questions center on how to manage and maintain a virtual environment. First, they need to be able to monitor and track virtual machines to know just where they are running. That, however, is just the first step. IT also has to know how well they are running, track resource utilization, and understand any failover implications in order to avoid issues regarding service level agreements pinned to the workloads executing on these virtual machines.

IBM Global Services can help to implement a critical set of tools for all of these tasks. The IBM Systems Director family of platform management solutions provides the means to discover and manage virtual and physical systems, storage, and networks from a single point of access. When IT is running servers at higher utilization rates, they need to be much more methodical about how they are managed.

The key management tools in the Systems Director family for virtualization are IBM® Director and its Virtualization Manager extension, which extends IBM Director’s deep physical knowledge of resources into the virtual world. IBM Virtualization Manager provides a systems administrator with a detailed understanding of the relationships between physical and virtual systems. With this software, a systems administrator is able to view and drill down on the relationships between resources, whether physical or virtual.

Application Profiling
The central tenet of “moving your mess for less” is to take everything in the physical environment “as is” into the virtual environment. While this is a workable plan, it is far from an optimal plan, especially for a large environment. Often the genesis for the “as is” plan comes out of the need to cut the growing cost of the electrical power and cooling needed to maintain racks of servers. With Moore’s law showing no signs of being repealed, new powerful servers, featuring multi-core processors, can easily be moved in to replace several older servers and simply reduce the environmental footprint of the datacenter by bringing it into an equal number of more efficient servers.

Nonetheless, any savings provided by a simple “as is” migration are dwarfed by what results when serious planning is undertaken to provide for optimal workload image reduction. This requires significant time and planning; but the payoff can be equally monumental. With a wealth of experience to leverage, IBM Global Services can take on the task of determining how best to implement virtualization in a site’s production environment. This can be an especially daunting project the first time it is attempted. That’s why many customers that conducted their own proof-of-concept studies turned to IBM Global Services to handle this phase of the project from start to finish.

IBM Services can help IT managers assess all of the site’s applications that have grown organically over time. Working with IBM Global Services, IT will often find more workload images running than were initially planned to meet the original reliability, availability, and scalability (RAS) criteria. It is not uncommon for a deep analysis to uncover upwards of 15 images running on older, less powerful servers. With new servers, both RAS and throughput metrics can be safely accomplished with just two or three images—a similar number of images to when the application was originally installed.

In such a case, moving all 15 images to run in virtual machines “as is” wastes more than IT infrastructure resources: It wastes expensive senior IT management time by tying up system and database administrators, who are far more costly. That wasted effort can also delay the implementation of new projects that could be contributing to the corporate bottom line. When all of the costs associated with resources and IT management are accounted for, the savings garnered from image consolidation can be eye-popping.

The Path to Service Management
To avoid wasting resources, time, and energy, a detailed examination of all of the applications in the site’s portfolio must be done, much like in a lifecycle management analysis for data and storage. This time, however, the analysis focuses on optimizing the number of workload images needed to maximize RAS and minimize overhead. Within that optimization process, savvy IT decision-makers will also discover the seeds for growing IT as a better service provider.

One of the first steps in reducing the number of workload images is a characterization of each workload in terms of its utilization of resources, including CPU, memory, and storage. The goal is to create a small number of workload categories. This leads to the adoption of a small number of standard server configurations capable of supporting various combinations of workloads.

From those finite sets of workloads and server configurations, it is now possible to create a set of service classes based on the specific levels of availability that IT can support and the resources that the application will consume. With a set of service levels defined, other departments will no longer come to IT with a preconceived collection of hardware, but with a request for a specific level of service and a description of the application’s characteristics. From that information, IT can propose a service level agreement and assign the application a standard server configuration. In this way, virtualization has helped IT begin thinking about itself in terms of being a service provider.

By working with IBM Global Services to analyze their unique business needs and current applications, companies can be sure that they are utilizing every benefit offered by a virtualized infrastructure. Both IT and the company bottom line will benefit from having to acquire fewer resources, which will be inherently easier to manage in a virtualized environment. More importantly, the extended benefits of a properly organized and streamlined IT environment can help facilitate governing IT as a service provider and that can facilitate further opportunities to implement lucrative line-of-business service projects.