Allocate, Grow, Change and Share Computing Capacity Without Disrupting the Application In today’s highly connected online world, high performance transaction processing applications play an increasingly central role in running the business. Because these applications serve end users in real time, service levels are an absolute priority. Slow application response times or downtime can translate directly into lower customer satisfaction, lost revenue or increased costs. Yet at the same time, these applications are often faced with highly variable workloads, driven both by unpredictable usage patterns and rapidly shifting business requirements. Traditional computing infrastructure is inadequate to run these applications because it is too rigid, slow to react to changing business requirements, and unable to cope efficiently with highly variable workloads. Over-provisioning helps, but having too many systems ‘just-in-case’ causes under-utilized assets that increase administration complexity, take up space, consume electricity, and require cooling. What if you had a pool of shared physical resources that you could govern via tailor made policies for each class of applications? Lots of flexible capacity at the ready for your applications, high availability, fewer moving parts, and higher utilization. Virtualization is a powerful vehicle to maximize service levels and minimize service costs by decoupling IT services from the physical resources on which they run. Virtualization enables a fundamental change in the way physical resources are deployed and delivered: from dedicated, rigid silos of capacity to pools of physical resources that are shared dynamically among applications and users. According to Gartner, “By 2010, a quarter of application demands in the ‘average’ fortune 500 business will be fulfilled from shared rather than dedicated sources.” ¹ In the server world, virtualization has materialized in two different technologies: OS-level virtualization and application-level virtualization. OS-level virtualization places an abstraction layer between the hardware and the OS, enabling multiple OS images to reside on a single physical server. Typical OS-level virtualization products include VMware™ ESX Server™ and Xen™. Application-level virtualization places an abstraction layer between the OS and the application, eliminating binary dependency and making the application portable across different categories of OS and servers. Typical application-level virtual machines include the Java™ virtual machine (JVM) and .NET Common Language Runtime (CLR). OS-Level Virtualization: Small Application Consolidation OS-level virtualization has become a prominent new trend in enterprise computing, with products such as VMware ESX Server and Xen gaining market momentum at a furious pace. The reason behind this success? Existing server infrastructure is only utilized at 10% to 15% of its capacity. Yet enterprises keep on adding servers faster and faster, and the associated capital, administration and facilities costs are spiraling out of control. Faced with this challenge, data centers have made it one of their highest priorities to consolidate applications, reduce the number of servers and increase the utilization of existing physical assets. OS-level virtualization has just what is needed: application consolidation without loss of OS, fault or resource isolation. ----------------------- ¹ Ben Ping and Donna Scott, "Positions 2005: Real-Time Infrastructure and IT Utility Redefine Delivery Models", Gartner Report, March 2005
Page 1 of 2 »
|