Making Peace in a Virtual World

Untitled Document

In many IT organizations, two camps have formed. The agility advocates want to unleash the power of virtualization technology, giving it free rein to flex the IT infrastructure for maximum agility. The control advocates want to rein in virtualization technology through control processes. They fear that lack of control opens the door to problems.

The two camps are in conflict. Agility advocates view control advocates as placing unnecessary restrictions on agility. Control advocates look at agility advocates as exposing the organization to unnecessary risk.

Both camps have valid concerns. Too much control hampers agility. Too much agility exposes the organization to risk. Striking the right balance is a difficult challenge.

Based on our experience working with many IT professionals who are tackling this challenge, we have come up with seven principles for a service management strategy in the virtualized world. By applying these principles you can find the optimum balance between control and agility. In so doing, you can bring about a peaceful solution that brings the two groups together. As a result, you can tap the unprecedented agility offered by virtualization, without exposing your organization to risk.

1. Embrace all types of change.
Changes occur through manual and automatic activities. Some changes are planned; some are not. Some are authorized; some are not.

  • Planned changes are anticipated and scheduled in advance. The manual provisioning of a new physical server and an update to a server operating system are good examples.
  • Unplanned changes may be anticipated, but they are not scheduled ahead of time. A manual fix to restore a failed server and automatic failover are examples of unplanned changes.
  • Authorized changes have all the proper approvals in place. Moreover, if they are manual changes, they are performed by a person authorized to make the change.
  • Unauthorized changes don't have proper approvals, or they are performed by someone who isn't authorized to make the change.

An effective service management approach accommodates all four combinations: unplanned/unauthorized, unplanned/authorized, planned/unauthorized, and planned/authorized.

2. Use only best-practice processes
IT Infrastructure Library (ITIL) guidelines outline best-practice change and configuration management processes. Control Objectives for Information and related Technology (COBIT) offers proven IT governance practices. Frameworks such as these, however, don't provide specifics on applying the practices to virtualization.

According to independent research firm IDC, "in companies adopting ITIL or other best practice frameworks, incorporating the virtual infrastructure into the process workflows will be crucial." By optimizing process workflows for the virtual environment, IT creates a single, integrated set of processes and controls to handle all changes, whether they are to physical or virtual resources.

That doesn't mean applying the same processes in the same way to every change. It means optimizing the processes for the type of change. For example, the level of control, auditing, and reporting required to add and provision a physical server differs substantially from that required for the automatic movement of a virtual server from one physical host to another.

By intelligently applying exactly the right level of control, auditing, and reporting to each change, IT can strike the optimum balance between control and agility, allowing the enterprise to exploit the agility offered by virtualization while minimizing risk.

3. Manage by exception.
To manage virtual environments effectively, IT needs to take a lesson from event management. Thousands of events occur in the IT infrastructure and it's impossible for the IT staff to intervene in every one. So, instead of flooding the operations console with alerts, well-architected event management systems manage by exception, presenting the IT staff with only those events that require human intervention.

As in event management, it is physically impossible -- and it's also unnecessary -- for the IT staff to get involved in every change in the virtualized environment. Many changes occur automatically based on pre-approved policies. The IT staff need focus on only the changes that require attention -- that is, the exceptions.

4. Maintain near-real-time visibility of the IT landscape.
Change happens quickly in virtualized environments. To work effectively, the IT staff must have a comprehensive, accurate, and current view of the IT infrastructure that reflects changes in near real time. This includes changes to virtual as well as physical resources.

Change information should be maintained in a configuration management database (CMDB) that shows the physical and logical relationships of all resources, and the services they support. For example, the CMDB should contain information on which virtual machines are supporting which applications and services, and which virtual machines are currently running on which physical host servers. The CMDB also maintains information on all planned changes and keeps a historical record of all changes made to each resource.

This comprehensive and near-real-time view enables the IT staff to determine the impact of a change on the overall IT infrastructure and on business service delivery. The record of all previous changes enables the staff to relate problems to recent changes, speeding problem resolution.

5. Automate and integrate wherever possible.
Automation enables IT to maintain control without hampering agility. It leverages solutions that automate change processes in the physical environment and optimizes these solutions for the virtualized environment. It also integrates the processes, out-of-the-box, with virtualization software control points for virtual machines.

Integration extends the sphere of control of the change management solutions to include changes made through the virtualization software -- whether these changes are performed automatically or done manually by the administrator. The result is a single change and configuration management system that encompasses all changes. As a result, the IT staff works with a single set of processes and supporting solutions, simplifying management and reducing training requirements.

6. Keep virtual server proliferation in check.

To control server sprawl in the physical environment, many IT organizations have established processes that govern the change and configuration of physical servers. Simply applying these processes to the virtual environment, however, is not the answer to preventing virtual server sprawl. IT must optimize these processes to govern the change and configuration of virtual machines. This ensures that all changes to both physical and virtual machines are performed with the right level of knowledge and control to eliminate both physical and virtual server sprawl.

7. Embrace heterogeneity.

Heterogeneity is a fact of life in today's data centers and will remain so well into the future. The typical data center includes physical servers that are hosting virtual machines and those that are not. It also includes disparate operating platforms, such as Windows, Linux, and UNIX servers, as well as mainframes. An effective service management approach accommodates this heterogeneity.

Peace at last

By applying these principles, you can bring the agility and control advocates together, bringing peace to the virtualized world. The benefits are well worth the effort. You'll contain the proliferation of virtual machines, preventing virtual server sprawl. You'll exert just the right level of control, auditing, and reporting to every change, ensuring that IT maintains continual compliance with corporate policies and government regulations -- all without limiting agility.

The optimum balance between control and agility enables IT to continually improve service. What's more, it reduces costs. IDC estimates that deploying ITIL-based process improvements for a virtual infrastructure translates into an average annual cost savings of $100,000 to $200,000.

In the end, IT moves faster and with greater flexibility without increasing risk. Everyone benefits. IT reduces costs and aligns more closely with the business, contributing greater business value. Business users gain the competitive edge of an agile and stable IT infrastructure that delivers the services they need, when and where they need them.

About the Author

Kia Behnia, chief corporate architect in the CTO Office at BMC Software, is responsible for leading the design of BMC Atrium, the industry-leading, service-enabling architecture for Business Service Management. He was previously CTO for the change and configuration products at BMC, and CTO for Marimba, Inc., which was acquired by BMC. Prior to joining Marimba, he served as a senior member of the technical team for Tivoli Systems, Inc. Behnia has more than 15 years of experience in the management of distributed systems and databases.

More by Kia Behnia