Top planning requirements for private cloud deployments
- 13 February, 2010 22:00
Virtualisation is understood as one of the key building blocks for private clouds. As a dynamic technology that enables IT organisations to reinvent how they think about management, it has the potential to make some things easier or make all things harder. Silo buying, heterogeneity, politics, poor integrations, and immature management tools can inhibit virtualisation's full value. IT leadership must address these challenges in order to set the stage for private cloud deployments, which many pundits forecast as the foundation for business growth in the next decade. Based on many customer conversations, we have compiled the top eight requirements for successful deployments of private clouds. These considerations enable IT to deliver business growth and contain costs, getting more from people, processes, and technologies.
1. Virtual Plus Physical Scalable Management
The integration of physical and virtual management is a key requirement to successfully deploying private clouds. Almost every IT organisation is utilizing virtualisation increasingly and starting to consider heterogeneous hypervisors. There is a growing need to have both "element" management and end-to-end "value" management that drive the need to use existing root-cause analytics, models-based management, integrated interfaces, and run book automation for both physical and virtual environments. Virtualisation does not obviate the need for management; it makes it more critical, especially as cloud computing is based on real time compute pools and dynamic allocation models.
2. Next Generation Architectures
Next generation data center architectures, such as Cisco Nexus virtual switches and the Cisco Unified Computing System will play a key role in enabling private clouds as they compress network, storage, and server connectivity into a single chassis. These architectures use virtualisation and offer customers new management paradigms for configuration, automation, security, and application-aware problem resolution. Management of these systems must enable customers to get more from the virtual layer, delivering dynamic capacity management and business service visibility.
3. Policy-Based Management
Management solutions must understand and aggregate polices from across different components; this is a notable requirement as private clouds dynamically shift and manage resources based on business demands. Pseudo standards such as OVF will be deployed over the next few years as the need to manage VMs from a service perspective (versus current component) enable service-oriented management. Policies will come in many different flavors, such as business or technology, and aggregation points will become critical to maintaining service levels in dynamic cloud environments. The integration of contractual agreements, chargeback, performance, Service Level Agreements (SLAs), and Quality of Service (QoS) metrics will be paramount in measuring value and delivering high quality IT capabilities that support and drive the business.
4. Process Standardisation
ITIL version 2 and 3, CoBit, and other capability maturity models will continue to garner attention as VM sprawl, compliance, availability, and security concerns increase the need for management and control. The fastest way to reduce costs is to standardize. Processes--both human and technology--offer opportunities to contain costs, not just reduce them across physical and virtual infrastructures. The integration of these environments becomes paramount.
5. Value Justification Analysis
IT and business leaders will be urged to define costs and make decisions based on business value, with supporting analysis for IT capabilities for in-house versus public cloud models. Each model should take into account the types of services and underlying cost structure to deliver those IT capabilities. Most important is the discussion of the business criticality of the IT capabilities and how strategic they are to achieving business growth. Value can be defined in many different ways; it's imperative that technology executives understand the business objectives and defend spending regardless if the IT capability will be delivered internally or from a third-party cloud provider. The definition of what the IT organisation can deliver, and at what cost, will become a central theme for decision making as cloud providers up the ante with low cost options for various IT capabilities.
6. Mainframe Virtualisation
The mainframe is very much alive with MIPS growing, as is virtualised mainframe environments. In some cases, enterprise customers are looking to integrate virtualised mainframe application processes with other types of physical architectures. As many applications traverse both mainframe and physical architectures, IT organisations are considering the impact on costs by looking at root cause analytics for application performance management. The ability to see a transaction as it crosses various pieces of the architecture, both physical and virtual, mainframe and distributed, is going to a key requirement for private cloud deployments.
7. Increased Automation
Automation is a key enabling technology of private clouds because it helps reduce and contain costs. Key standardized processes, such as change and configuration; application release management; financial management, with a migration towards service transparency; service operations; and related ITIL definitions offer tremendous automation opportunities. Physical and virtual processes must be integrated, but this is not enough. Automatic problem identification, thresholding, and problem remediation are key capabilities that enable private cloud deployments.
The idea of "self-service" is exploding with customers as they consider what it means to them. The notion is simple, to empower employees to request compute resources on demand and deliver those resources automatically. Private cloud deployments increase the value of self-service as customer demands will require a linkage to chargeback models, resource availability, network, storage, and server knowledge, and an easy to use portal that casual IT and business workers can use.
Recent conversations with customers show just how far we have come, and provide a glimpse into where we are going. One financial service firm wants to have a "branch in a box" with all the key applications and infrastructure virtualised, and a layer of management to provide control and visibility. Another customer is interested in having a dynamic fabric and automated assembly line for business applications. Yet another wants an automated, self-service cloud for their development teams. IT organisations are adapting to the speed and pressure to deliver private cloud services across their infrastructure fabrics. The net benefits of these capabilities are clear: reduce costs, contain costs, drive tighter business impact, and get more out of people, process, and technologies.
Stephen Elliot is vice president of strategy for CA's Virtualisation and Service Automation business unit.