Whatever the name – Grid Computing, Utility Computing, the Adaptive Enterprise, the Agile Enterprise, Capacity On Demand – the concept is very attractive to organizations that seek to be able to seamlessly adjust and allocate resources to pursue business opportunities as the moment they arise.
The goal of these concepts is to synchronize IT resources to match fluctuating service requirements, using automation and virtualization technologies to provide transparent resource support and reduce waste, which occurs when server resources are over-provisioned. The resulting benefits that can be attained are universally appealing to today's businesses.
- Increased efficiency and lowered costs by consolidating and optimizing application performance
- Minimized business service risk by assuring 100 percent availability and service levels
- Increased business agility at optimal cost by automatically ensuring appropriate capacity "just-in-time"
Although the principle of concepts like Utility Computing and Capacity On Demand are extremely basic, until now it has been possible to attain only on large-scale proprietary systems such as mainframe-based environments. Across distributed computing environments such as Unix, Windows and more recently Linux, over-provisioning has been a major source of waste for decades, from the client/server era to today's commodity computing. There are three distinct reasons: Organizational structure, ownership and control, the overshadowing risk of being caught with insufficient capacity; both coupled with the perception that the incremental hardware capacity involved was "cheap". However, while the impact of lost opportunity is significant, the extraordinary financial and opportunity costs that result from sustained levels of excess capacity are greater and largely unnecessary.
The costly over-provisioning of server resources that has been ingrained in IT management for years can be readily dissolved by adhering to the principles of Service-Oriented Resource Management (SORM). As the name suggests, SORM involves managing IT resources to meet varying service level requirements, instead of reflexively over-provisioning to provide vast levels of excess resources, "just in case" they are required. SORM enables resource levels to finally be synchronized with demand and provides a modular, open, and pragmatic approach.
Real-time resources: A concept ahead of its time?
To achieve the full benefits of SORM requires managing IT resources as a service and catalyst for business value creation. This calls for organizations to function as a service provider, understand costs, guarantee service level agreements, integrate processes, and provide dynamic capacity management. Currently, less than 10 percent of all enterprises are functioning at this level.
Although these capabilities appear to be straightforward and attainable, it should be considered a futuristic IT goal. Across the enterprise, individual departments have a deeply ingrained culture of owning servers and applications, and most are reluctant to relinquish it. Furthermore, the departmental orientation of IT asset ownership and operation is counter-intuitive to establishing effective cost allocation methods for virtualized resources. Finally, most enterprises lack a standardized, normalized way of allocating costs for the use of a virtualized and shared infrastructure.
Service-oriented resource management delivers business benefits
SORM allows enterprises to begin achieving this futuristic goal by incrementally creating the virtual datacenter, an IT environment in which server, storage, and network capacity can be automatically re-allocated from another application area that has excess capacity. The virtual datacenter keeps resources balanced according to business policy, and when a component fails, or additional capacity is required, automatically allocates the appropriate spare, underutilized or lower business priority resource to meet the need.
The ultimate goal of server and datacenter virtualization is to simultaneously increase business agility. This significantly reduces the chronic over-provisioning of server resources, which in large enterprises routinely creates a five-10x surplus of computing capacity. SORM principles can enable significant operating cost reductions – easily 40 percent or more – in servers and their associated costs, while ensuring that all business services are provided the resources they need and mitigating the risks associated with unpredictable demand. By standardizing server management practices, IT departments can create repeatable solutions that free up capital wasted on the maintenance of excess computing capacity.
Business risk mitigation is another major benefit that can be derived from proactive resource management and datacenter virtualization. In today's complicated IT environments, service demand and application environment change is too dynamic to predict. Despite decades of advances in service-level management, systemic over-provisioning persists. The alternative – potentially insufficient computing capacity at peak demand times – exposes companies to substantial business risk.
By adopting a strategy such as SORM, that ensures that resources needed to meet unpredicted demand peaks are available at the right time, organizations dramatically reduce risks and can maintain high service levels. Prioritized resource allocation from large, but shared, pools of excess capacity ensures always having enough capacity. Additionally, as the virtual datacenter is automated through SORM, less human intervention is required, further reducing business risk.
In conclusion, many enterprises are working towards providing optimal service levels at a justifiable cost. Companies can take a giant step toward reducing the costly over-provisioning of sever resources with the principles put forth by SORM, a methodology that manages IT resources to meet varying service level requirements. By adhering to open, standards-based SORM methodology, companies can be put on a path toward achieving real-time enterprise infrastructure.
Dave Wagner is a solutions management director for Capacity Management and Provisioning at BMC Software. He is responsible for driving overall solution strategy, pricing, requirements and positioning of BMC's families of proactive performance analysis, modeling and dynamic provisioning and orchestration solutions across leading enterprise platforms and their associated application environments.