Q&A: Creating a Virtualization Strategy

Before embarking on a virtualization project, you need two things: a good plan and the best implementation strategy. We offer tips for building both.

If you're exploring how to get the most out of virtualization in your data center, you need a plan and you need to know how to avoid the common mistakes your colleagues have made when implementing a virtualization plan.

We asked Scott Farrand, vice president, Infrastructure Software and BladeSystem at HP about the key elements of a virtualization strategy, what benefits you can expect, how to avoid common mistakes, and how to meet the technical challenges of a heterogeneous environment.

Enterprise Strategies: How do customers assess what type of virtualization (server, storage, networking, etc.) is appropriate for their business/IT needs?

Scott Farrand: Customers should start with the business goals they are trying to achieve -- e.g., are they primarily looking for short-term cost savings or are they also interested in benefits such as an increased ability to respond to change and reduced risk? Although short-term benefits can be achieved by focusing on a single area such as server virtualization, the longer-term benefits require a broader strategy across servers, storage, and networking.

The most important step to making effective virtualization design decisions is to first assess the current data center resources. Customers should evaluate their specific server, storage, networking, and management capabilities to determine when and where it makes sense to virtualize.

At the end of the day, each project requires careful, individualized assessment of technology resources as well as business requirements.

Why should an enterprise create a virtualization strategy, and what are the key elements of such a plan?

Implementing a virtual environment requires thorough planning and can become overwhelming unless administrators have the proper knowledge and tools available. Customers should outline a strategy at the outset of a virtualization project to ensure that they will achieve maximum benefits from the deployment.

Key elements of a plan should address:

  • Data center assessment: Which IT assets are siloed and underutilized? Capacity planning tools can help address this question and provide useful insight.
  • Management: How will I manage my data center now that I’ve introduced another layer to my environment by deploying virtual machines? Some of the barriers to adoption can be eliminated from the outset by considering management options at the start of the planning process.
  • Ownership: Who will ultimately be the virtualization champion in your organization? By assigning ownership to one person, an integrated approach is fostered more easily.
  • Specialized knowledge: Vendor and other third-party services provide skillful expertise when IT staff has limited experience. Services representatives help businesses develop a specialized virtualization strategy designed to meet their individual needs. Customers can also benefit from implementation, management, and staff training services.

What are some of the benefits IT expects of virtualization? Are these benefits being realized?

Reducing costs is often the primary motivation for virtualization projects. By improving resource utilization and efficiency, companies reduce the need for physical assets. Generally, most companies are successful in reducing costs for spending areas such as servers, floor space, and power. For example, virtualization is being used to create more power-efficient facilities by decreasing the number of physical assets required to run the technology and cool the data center, further cutting costs.

Beyond costs, virtualization provides a technology foundation to deploy applications or business services more quickly, which accelerates business growth. A virtual infrastructure provides resources where and when needed with minimal or no operational disruption -- keeping the business running 24x7 for a 24x7 world. Tapping virtualization’s potential here, however, requires companies to think beyond virtualization technologies to the processes they use to govern and managing their data centers. The payoff is higher, as long as they make the right investments.

What role does risk play in deploying virtualization technologies? How can customers alleviate risk in their data center?

Virtualization brings its own unique set of technical challenges that can lead to security and compliance implications if left unaddressed. Potential risks include malicious software in one guest operating system attacking another operating system; overlooking the security needs of virtual or physical resources with regards to software configuration, updates, and patches; auditing and performance monitoring; and neglecting compliance requirements for a guest operating system.

Planning a management strategy is also required to minimize risk. Consolidating workloads across a virtualized data center includes movement and migration of the workloads. Tools for effective movement and capacity planning are required to ensure the cost savings of a virtualized data center.

When confronting the issue of virtualization security, the challenge is to avoid creating policies that prove too restrictive. Customers need to question if all of the benefits of virtualization are being achieved with their current security policies. The ideal solution ensures that users retain control of their infrastructure by ensuring virtualization is not bypassing existing security controls.

Beyond the technical challenges are governance and process risks. For example, when new virtual servers can be created in a matter of minutes, storage provisioning processes may suddenly appear to slow. Without proactively addressing these challenges, companies can find themselves unable to take full advantage of the technology’s potential.

How do enterprises ready their infrastructure for virtualization?

There are several ways customers can prepare their infrastructure to manage virtual machines in order to fully realize the benefits, including:

  • Create a balanced architecture with hardware that has comparable performance capabilities. Consider all aspects of your infrastructure to maximize the potential of each component -- for example, the full potential of quad-core servers can be realized only when paired with equally impressive memory and networking capacity; otherwise, performance benefits will be compromised.
  • Avoid a multi-model setup by migrating to a bladed infrastructure with shared storage. Most applications have specific recipes for the amount of network, storage, and server resources they need, which can make it difficult to modify components to accommodate virtualization. Converging on a bladed infrastructure with a shared storage model allows customers to fully realize the flexibility benefits offered by virtualization.
  • Utilize a single, large network connection to flexibly distribute network bandwidth as needed. A single high-bandwidth network connection with virtualized connections to servers and storage allocates bandwidth as required and enables quick configuration changes within the pipe, without moving or unplugging anything.

What common mistakes do IT departments make when implementing virtualization in their environment? IT departments often begin virtualization projects on an ad hoc project basis. Before long, thousands of virtual servers flood the environment, while businesses end up with two sets of tools to manage two separate infrastructures -- the physical and the virtual. This can mitigate the potential savings of initial virtualization projects and defeat the goal of the project in the first place.

Another common mistake can be to overlook cultural barriers that should be anticipated when deploying virtualization technologies. Virtualization changes the data center and significantly impacts the organizational structure, as well. The sharing of IT resources that virtualization enables within the data center requires that the lines of business (or departments or application teams) learn to share as well. This can mean, for example, that departments are forced to share machines that they previously owned by themselves.

This is a significant shift in the standard work paradigm that IT staff is accustomed to and raises questions that need to be addressed, such as: Who will have control? What new skills will be required? Will my IT department need to downsize?

What best practices can you recommend to avoid these mistakes? A carefully planned strategy is key to addressing such issues as management of virtual and physical resources before the first virtualization project is ever initiated. The plan should also anticipate obstacles associated with cultural changes stemming from virtualization, putting as much emphasis on people and processes as it does on the technology.

How is managing a virtual environment different from managing a physical environment?

Virtual environments bring a unique set of issues that require close management to minimize security issues, performance problems, or server sprawl. Because it is very easy to create new virtual machines from each physical server, server sprawl can rapidly ensue and lead to 50 virtual machines to manage.

Prioritizing management will ensure that both physical and virtual machines are accounted for while reducing the time spent on attending to virtual machines. In turn, staff efficiency increases and service levels improve so customers fully reap the benefits of virtualization.

Virtualization can also transform standard IT processes such as provisioning. What is typically a manual process taking weeks or months can be automated down to minutes or days with the right management software.

Data centers are usually heterogeneous environments, which presents a challenge to IT. Does virtualization make management across heterogeneous environments easier or harder?

Given that most environments are heterogeneous, it’s important to ensure that the tools you choose to manage the environment apply to both physical and virtual, as well as multi-vendor environments.

Instead of having separate management consoles for physical resources and virtual resources, choose management software that views them the same way and lets you manage them together -- at anytime from anywhere. In this way, the challenges of managing across heterogeneous environments are greatly alleviated.

How does virtualizing the network connections create an integrated network of resources across the data center that impacts service delivery?

Virtual connections, used in place of conventional switch modules, abstract and pool the server-edge connections to allow server administrators to independently manage server blades and their connectivity. When combined with a holistic management tool, customers can easily build and save automated workflows to remove the cost and time of manual processes.

With virtualized network connections, resources are assigned to requests as needed, and then returned to the pool once the service requirements are completed, thereby optimizing utilization of the infrastructure. In turn, the entire life cycle of the infrastructure is automated, from provisioning through retirement -- fundamentally changing the way technology is used to deliver business services, while reducing overall infrastructure costs.

What products or services does HP offer to help with the issues we’ve discussed?

HP offers the broadest range of virtualization products and services available, enabling companies to virtualize with success.

HP recently unveiled a system we’re really excited about -- the HP BladeSystem Matrix (http://www.hp.com/go/matrix) a convergence of software, server, storage, and a networking platform that automates service delivery for the data center. Matrix can create a pool of resources for hundreds of diverse applications by integrating virtualized assets with a single management interface.

Matrix comes with HP’s Virtual Connect Flex-10, using interconnect technology that eliminates additional network equipment and costs.