Four Priorities to Enhance the Virtualized Infrastructure
Four Priorities to Enhance the Virtualized InfrastructureHeres a question every CIO will find familiar: How can I get more business value fromthe data center?Virtualization is certainly a step in the right direction. Virtualization can make the datacenter more responsive to changing conditions and more scalable and flexible in thepursuit of business goals—in short, more agile. Meanwhile, if implemented effectively,virtualization can also significantly reduce costs. This combination is bound to please ourhypothetical CIO.However, there is a difference between any given abstract idea and its bestimplementation. And for organizations interested in taking advantage of virtualization tocreate more business value, there are certainly better and worse ways to go about it.Because each organization has a unique context—characterized by unique assets,processes, strengths, weaknesses and business risks—each will require a unique strategyto leverage virtualization to best effect. A CIO is chartered with developing that strategyand adjusting it to fit the organizations needs as closely as possible.IBM can help. Today, IBM offers a new, straightforward, logical framework, comprisedof four priorities, that todays CIO can use to create just such a strategy of virtualization.Each priority in the framework is independent of the others—basically, a different lensthrough which the CIO can perceive and improve virtualization in the data center. Somepriorities may not require as much time or investment as others to fulfill; depending onwhat an organization has already accomplished, some may not apply at all.Furthermore, the order in which they are presented here is merely a suggestion. Whilemany organizations will get excellent results via this sequence, they can actually bepursued in any combination. The specific sequence should be chosen by the organizationas determined by its needs. Yet all four can help drive down costs, drive up performance,and enhance business agility.And together, they help ensure that the benefits of a virtualized infrastructure are notmerely an abstract possibility, but an operational reality—a reality thats been tailored forthe best business outcome.Priority: Consolidate resourcesThis priority is probably the most familiar to today’s organizations because it reflects afamiliar description of most of todays data centers: thousands of relatively low-endsystems, each of which is utilized less than twenty percent of the time, but each of whichcontinually draws power and generates heat.It stands to reason that fewer, higher-end systems, utilized better, will deliver a betterbusiness outcome. Such an architecture makes the data center less physically complex,
increasing valuable floor space, while also minimizing the possible points of failure. Thedirect result: by consolidating systems, IT will become more efficient and operating costswill fall.Virtualization is the key technology needed to consolidate systems in this manner. Givenhigh-end hosts, capable of simultaneously running dozens or even hundreds of virtualservers, the overall utilization of each host will be much higher than before—in manycases, exceeding fifty percent. Yet the energy costs will be much lower. And becauseeach high-end system generates less heat than the equivalent number of low-end systems,there is less heat to dissipate, translating into lower cooling costs as well.Furthermore, the extensive physical resources of each host (processing power, memory,or storage) can be allocated among those virtual servers flexibly, in proportion tochanging business requirements. This means that application and service performancecan scale up to meet demand (or scale back down when demand falls to reduce costs),instead of being restricted by the minimal physical resources available on a low-end host.Priority: Manage workloadsAnother major priority in the IBM framework is workload management.The ideal workload implementation would be this: all workloads are fulfilled, to alltargeted service levels, under even the most unpredictable and demanding conditions—allwhile using the fewest resources. However, even given a virtualized infrastructure, thatoutcome can be exceptionally difficult to achieve.Part of the problem is that the management challenge will often increase throughvirtualization. Where one application or service was previously typically supported byone physical host—a relatively simple paradigm—that is no longer the case. Now, theremay be hundreds of virtual servers per host, and there may be hundreds of system imagesassociated with those virtual servers.Managers may need more time and energy to track problems to root causes, or implementnew strategies, than they did before. The significantly increased technological complexitywill thus unfortunately generate unwanted costs over time, threatening to diminish thepotential business benefits promised by virtualization.Workload management, then, is another priority to consider. Organizations will need tosimplify the overall management challenge via best-in-class, centralized tools that arespecifically designed for a virtual infrastructure, allowing managers easily to track statuslevels, perform everyday tasks, and implement any required form of change previouslyall under a single pane of glass.Tools of this type, in fact, have the effect of abstracting out technical complexity. Withtheir help, managers can shift the focus from the technology per se to the businessoutcome of the technology: the extent to which the infrastructure is fulfilling workloadsand hitting business targets.
And if a given service is not performing up to dynamically-changing requirements, moreresources can be allocated to it, quickly and fluidly. The result will be improved customersatisfaction (for external services) or user productivity (for internal services), both ofwhich are crucial to overall business success.Priority: Automate processesWhat will be the swiftest, least-expensive response to changing conditions in avirtualized infrastructure?In many cases, particularly those of an everyday nature, it will be an automated response.Automation, where applied in a suitable context, delivers a far lower operational cost, afar faster response time, and a much more consistent response by eliminating the need forhuman oversight. While complex, sophisticated tasks will certainly require a dedicatedhuman intelligence, they should be the exception, not the rule.As a simple example, consider the sequence of events involved in provisioning a virtualserver with an appropriate stack of software: OS, applications, middleware, drivers, data,and other elements. An automatic provisioning process will be dramatically faster than amanual provisioning process. It will also be completely consistent from case to case. Thismeans that any services or applications that rely on that virtual server will not fail as aresult of inadvertent human errors in configuration that might otherwise come into play.Automation can enhance a virtualized infrastructure in many such respects. The terms ofservice level agreements that specify ongoing performance levels, for instance, can moreeasily be met if they are dynamically fulfilled through automation. Achieving compliancewith government mandates will be simpler and less expensive. Crucial processes such asdisaster recovery, involving predictable actions based on known resources, will certainlybe faster and more complete if automated, meaning the organization can minimize thebusiness impact of the disaster.In short, automation can help make virtualized infrastructures more efficient, cost-efficient, responsive, and available—even self-healing, in the event of everydayproblems.Priority: Optimize service deliveryAt the leading edge of virtualized infrastructures are those which have been trulyoptimized for business goals—usually by putting business users in control as directly aspossible.Consider this situation: a line-of-business manager, with no IT background, conceives ofa new service that would help his team (or that is needed by external clients/customers).Rather than submit a request for the service through IT, this manager can insteadcapitalize on an existing, highly-optimized virtual infrastructure to create the servicedirectly, over the Web, using any standard browser.
This compelling idea will certainly require an advanced virtualized architecture that hasbeen developed and optimized to facilitate it. One example of such an architecture: cloudcomputing. Given a cloud, the time required to translate service ideas into actual, up-and-running services can fall dramatically from weeks to hours. In parallel will fall costs:management costs, operating costs, and even capital expenditure costs.The result will be tremendous business agility with a relatively low price tag—both todayand tomorrow.Additional Information:Bullets:• Getting best results from virtualization will mean developing a tailored strategy• IBM offers a new framework of four virtualization priorities designed to simplify thatprocess• These four priorities are: Consolidate Resources, Manage Workloads, AutomateServices, Optimize Delivery• They can be pursued in any order or depth the organization requiresTalking points: see aboveTweet tagline: Tailor your virtual infrastructure to your needs via IBMs priorityframework: Consolidate Resources, Manage Workloads, Automate Processes, OptimizeDelivery.Abstract: For organizations looking to get more business value from virtualization,IBMs new, modular framework of priorities can help—regardless of the infrastructurecurrently in place. These four priorities—Consolidate Resources, Manage Workloads,Automate Processes, and Optimize delivery—can be pursued at any necessary level ofdepth, and in any sequence best suited to the organizations needs.Tags/keywords: virtualization, priorities, consolidate, workloads, automate, optimize,cloud computing, cloud, reduce costs.