Private Clouds: 5 Considerations for Optimizing the Dynamic Data Center
If you’re moving to private clouds, keep these considerations in mind as you plan your migration.
By Jason Cowie, Vice President of Product Management, Embotics
The concept of the private cloud has been around ever since the first data centers were formed. With the advent of virtualization technology, this idea has moved beyond the realm of vision and become possible for all data centers (large and small).
Several common challenges to building and managing private clouds affect most organizations. These include changes to IT organizations, scaling with limited resources, and managing complex environments. Enterprises can overcome these issues, but not with rip-and-replace methodologies. A more pragmatic path can help data centers evolve from their current, limited use of virtualization towards truly dynamic environments and private clouds.
Along that path are five key considerations every organization must face.
1. Virtualization requires fundamental changes to IT organizations.
When a data center contains a low percentage of virtualized servers, virtualization can be managed as just another IT group. That quickly changes as the percentage of VMs or virtualized IT services increases. The reality is that server virtualization for the masses requires innovative approaches to management. It adds significant management challenges, and perhaps more importantly, requires fundamental changes to IT organizations and operations.
As long as the focus of virtualization is server consolidation (i.e., a low percentage of physical servers involved, few production VMs, and a minimum number of mission-critical applications), the complexity is low, and one team can manage it quite comfortably.
However, as enterprises scale their virtual environments and run more business-critical applications on virtual servers, complexity increases, as does the workload on the virtualization staff and the pressure to build and maintain reliable supporting infrastructure.
To get to the private cloud, IT organizations will have to reinvent themselves, aligning security, operations, application, server, storage and network teams towards common goals and service level agreements. They’ll also have to implement a new, integrated data center management system, uniquely designed for the virtual environment. Furthermore, they need to do this while still dealing with other important challenges, including day-to-day administration, cross-functional coordination, and education.
2. Automation helps alleviate the resource challenge.
When members of the Network World Tech Connections Research Panel were recently asked to choose their top three challenges with respect to virtualization, the first item on the list was scaling up with limited IT staff and resources.
There are many tasks inherent in properly managing IT infrastructure components, beginning with the virtual and physical servers themselves. Although virtualization does reduce the number of physical servers, it also typically increases the number of server instances, all of which require administrative time and systems to manage server operating systems, applications, and data.
VMware’s vSphere management console is somewhat effective, but most organizations find that as the environment scales, the time spent on regular daily and manual tasks becomes all-consuming for the administration team.
Adding staff is not a practical solution for most organizations. Instead, they need supplementary management, automation, and reporting that can standardize the environment and allow IT teams to manage the virtual infrastructure more effectively and proactively. By automating manual tasks such as asset tracking, provisioning, decommissioning, and capacity planning and alerting, organizations can focus on more strategic priorities that improve the quality of service being delivered and the responsiveness of IT.
3. Poor management of complex environments will lead to virtual stall.
As long as virtualization initiatives are small and focus on simplistic workloads, they are typically easy to manage within the operations team. However, as the use of virtualization increases, the combination of the increased volume of virtual servers and the increasingly critical nature of the applications means that other management silos in the data center need to get involved.
Many analysts view virtualization as a technology that disrupts the status quo of the traditional data center organization and legacy management systems. The deployment of IT services in a virtual world still requires the coordination and collaboration of teams spanning operations, security, storage, network, servers, and applications. Unfortunately, legacy management, automation, processes, technology, and approaches are, more often than not, inadequate for the virtual world.
4. When you can’t leap forward all at once, take a few steps in the right direction.
Ask any change specialist what is required for organizational evolution, and you’ll learn it’s a combination of three things: a vision of the future, dissatisfaction with the status quo, and the first few steps. The first two elements certainly exist in evolving virtualized data centers. The final part -- initial steps forward -- should permit organizations to take their time to ensure they have the right solution in place while maintaining momentum and proceeding in a controlled and planned manner until they reach the goal of the private cloud. Those steps might include breaking complex projects into smaller components, deciding which management features are needed to start, and establishing a management platform that can grow over time.
5. Choose the right technology solutions to support your private cloud strategy.
VM management is essential in progressing toward a pragmatic private cloud. Look for solutions that include discovery and reporting, capacity planning, self-service options, life-cycle management, change and configuration management, resource optimization, and policy automation.
Start by implementing your reporting and monitoring “pane of glass” to provide oversight over the whole environment. Get line-of-business buy-in with a self-service portal, and resolve immediate problems with management features such as capacity planning. As needed, add workflow and policy-based automation that will deliver long-term solutions rather than temporary fixes.
Choose a complete, fully integrated product that eliminates the need to learn, operate and integrate disparate tools while allowing you to easily identify and prioritize problem areas. An intuitive user interface and easy integration with other systems are also essential.
The Pragmatic Path to the Private Cloud
At its core, a private cloud is little more than a virtualization infrastructure that combines automated management capabilities with business process integration to deliver IT services on demand. These three elements sound simple, but they represent a significant change for any organization and require both careful planning and time.
Of these three ingredients, virtualization technology is probably the easiest part of the private cloud to achieve. Keep in mind, though, that an effective private-cloud strategy incorporates more than just the infrastructure implementation. Equally important as the technology is the management and process integration. It is these two pieces that resolve the resource and complexity challenges that cause so many virtualization deployments to stall out before they reach their potential.
The notion of a private cloud is no longer hypothetical. Getting to the cloud is possible, but it will take an acknowledgement of the complexities involved, a willingness to proceed, adequate planning, adherence to best practices, and the right management system for the job.
Jason Cowie is the vice president of product management for Embotics where he oversees product direction and strategy. Previously, Jason was the general manager at EMC where he was responsible for the server management business and played a key role in the acquisition of Configuresoft. You can contact the author at email@example.com