Businesses are demanding that their data centers support growth and agility, improve availability, minimize ongoing support and management costs, and reduce power and cooling demands. These seemingly conflicting requirements are forcing IT managers to take a fresh look at the workhorse of the network infrastructure — the server.
Many organizations rely on a combination of low-end commodity servers and high-end symmetric multiprocessing (SMP) servers or legacy mainframes. Traditionally, enterprise-class servers have been used to support core business applications while commodity servers were utilized for departmental applications and tasks such as Web hosting in which the workload can be divided among multiple machines. Low-end servers and clustering technologies deliver the functionality and performance to support “scale-out” configurations in which commodity servers are added as workloads increase.
The trend toward scale-out deployments began during the economic downturn of 2001-2003, when IT managers were forced to slash capital expenditures. Commodity servers enabled them to purchase incremental server capacity as needed at a lower initial cost. Now, however, organizations are finding that the scale-out approach taxes data center space, power and cooling loads, and inhibits the organization’s ability to respond to changing business requirements.
“Organizations can no longer afford to throw a bunch of low-end servers at the problem. Although the capital cost of x86 servers continues to drop, operational expenses are skyrocketing as organizations struggle to manage, support and power hundreds or thousands of distributed commodity servers,” said Steve Rogers, VP of Engineering and Strategic Planning, Jeskell. “IBM anticipated the demand for x86 servers that deliver the reliability, manageability and performance of traditional scale-up architectures. With System x, IBM has created an x86-based solution that enables customers to reduce IT overhead and gain technical, business and ‘green’ benefits.”
It All Adds Up
The proliferation of low-end file, print and application servers and their associated storage devices has created a complex and difficult-to-manage computing environment that consumes too much energy and creates security concerns. IBM is helping customers solve this problem with a scale-up strategy based upon industry-standard components and IBM innovation.
With decades of experience in mainframes and supercomputers, IBM is uniquely positioned to create high-performance platforms with a balanced system design that minimizes bottlenecks. IBM System x servers fully utilize today’s multi-core x86 processors and faster memory to accommodate rapid growth in application demand and support changing business requirements.
For example, the IBM System x3455 and System x3755 servers featuring Quad-Core AMD Opteron processors offer exceptional power efficiency, performance and virtualization capabilities for today's most demanding data centers. The IBM System x3455 delivers robust performance for scientific, technical and financial applications with IBM's Xcelerated Memory Technology, which eliminates traditional high-performance computing application bottlenecks by delivering faster access to data. The System x3755 has a flexible design that allows customers to increase the size of their system as their databases, virtualization environments and collaboration applications grow. The x3755 provides a three-socket configuration that outperforms many competitive four-socket configurations.
IBM also offers BladeCenter solutions based upon Quad-Core AMD Opteron processors. The IBM BladeCenter LS22 is a two-socket blade server ideal for companies that run high-performance applications. Customers can purchase the LS22 with a memory booster that gives it 96 percent faster memory for improved application performance. The IBM BladeCenter LS42 consumes up to 16 percent less power than comparable blade offerings. It also provides 20 percent faster memory, so it is well-suited for virtualization, databases and other memory-intensive applications.
“With the 2+2 program, customers can start with two sockets on the BladeCenter LS42 and System x3755 servers and expand to a four-socket server as their business needs demand,” said Rogers. “This allows companies to pay only for the capacity they need now, giving them the flexibility to scale their server while protecting their investments.”
Combine and Conquer
Improved reliability, availability and serviceability have long been a benefit of enterprise-class servers, with sophisticated failover capabilities that help maintain data integrity and business continuity during planned and unplanned downtime. These features are especially important in virtualized environments, in which a single system supporting multiple applications.
“Because multi-core processors make even low-end servers suitable for virtualization, IBM has made high-availability a part of every System x solution,” said Rogers. “System x servers also feature innovative power and cooling designs and built-in virtualization optimization and management features. With higher utilization and better control over workloads, customers can delay server purchase and better manage power usage, while improving availability through redundancy.”
Scalability — the ultimate strength of enterprise-class servers — is the Achilles’ heel of low-end servers. When the processing capacity of these servers is reached, more servers must be added, increasing data center headaches. The System x iDataPlex solution helps relieve massive, sprawled-out data centers that are pushing the limits of power and space available to them. Designed to dramatically increase compute density while significantly decreasing energy consumption and associated data center costs, iDataPlex uses 40 percent less power and increases by five times the amount of computing done in the same space. iDataPlex also uses industry-standard components as well as open source software such as Linux to lower costs.
A scale-up strategy is proven to simplify an organization’s IT infrastructure, enable greater control of computing resources and lower TCO by reducing infrastructure complexity, streamlining management, and reducing power and cooling costs. With the IBM System x architecture, organizations can achieve these results with unique features that transform the x86 platform into enterprise-class servers.