The incentive for refreshing your servers is being able to do more to support the everyday and mission-critical goals of the organization. Spurred by new technologies and virtualization, servers are running faster processors with greater numbers of cores and threads per processor, thereby enabling greater computing capabilities. Yet server refresh expenditures can sometimes be difficult to justify. 

By replacing older, less capable servers with newer, more powerful systems, a business can perform the same amount of computing with only a fraction of the total number of systems. Every three years, with increasing computing power, it makes sense to replace the old servers and modernize. This is especially true as your workloads grow and you’re trying to lower your management, power and cooling costs.

Reasons for refreshing servers

Productivity – IT departments are pushed to increase performance and business capabilities, often without additional funds. Each new generation of servers delivers baseline performance gains between 15 and 40 percent. The latest servers can address higher volumes of memory than previous generations. The more data you can put in memory, the faster that data can be processed. Moving to a new server architecture to reduce processing latencies can be particularly important for performance-intensive industries, such as financial services, where the speed of trades and other time-sensitive functions drive business.

Operating system advancements – In many cases, an older server cannot support the latest operating systems from Microsoft, VMware and Linux. The latest operating systems have new capabilities or improvements with virtualization, reliability, security, storage handling, provisioning, server management and licensing. The new operating systems also take advantage of the latest Intel processor features.

Power and cooling – Moving from one server generation to the next typically results in a significant savings in power consumption. The less power consumed by servers means less heat generated and lower losses for power distribution. This also reduces the cost of heat extraction. Lowering server power consumption, along with controlling that consumption, may allow the deployment of additional servers per rack while keeping within an overall power budget. In many cases, between two generations of servers, you can nearly double compute performance using an equivalent amount of wattage.

Consolidation – Fewer machines mean smaller numbers of boxes to manage, maintain and physically house in expensive data center real estate. When one new server does the work of two or three older servers, the environment becomes easier to manage. An additional benefit to a smaller server footprint is a reduction in the number of maintenance contracts.

Business operations – New servers with increased processing capability, improved storage, faster I/O and increased memory enable client applications and processes to run faster and more efficiently, and overcome workloads problems. This allows company processes to run smother, handle big-data analytics and react quicker to ever-changing customer demands.

There's a great IDC paper that goes into detail on the the cost of retaining older IT infrastructure. The paper looks at how the buy-and-hold strategy actually adds cost to the data center in the form of increased hardware maintenance costs, lower energy efficiency and security vulnerabilities. IDC found that as servers age beyond the three-to-three-and-a-half year optimal replacement cycle, annual maintenance increases significantly. They concluded that upgrading data center servers resulted in an ROI of 150 percent over three years with payback on the initial investment of 11.7 months.

The refresh business case

Investigate these areas to help justify the server refresh business case.

Operational cost reductions – Power and space efficiencies from investing in newer-generation servers become key points to focus on within a business case.

• Data center and real estate costs
• Heating, ventilation and air conditioning (HVAC) costs
• Total data center power consumption
• Connectivity costs
• Power consumption of older inefficient servers
• Management tools and people costs

Maintenance cost avoidance – Any server eventually reaches end of life (EOL), and waiting until EOL puts infrastructure teams in a risky situation. During that time, a business depends on unsupported hardware that lacks readily available replacement components. As an EOL milestone date approaches, research whether a server refresh outweighs the potential investment in extended maintenance support.

Service reliability – In the current, round-the-clock service-availability climate, businesses increasingly rely on IT to host applications and services on a solid, reliable IT infrastructure. But greater demand for application availability in turn puts greater demand on the physical server platform.

Consolidation – Many organizations now refresh hardware to gain server consolidation. These benefits offer the opportunity to virtualize more resource-intensive workloads that were previously unsuitable for virtualized environments.

When proposing a server refresh business case, include the financial benefits of higher virtualization ratios. Highlight to stakeholders that this continued, aggressive consolidation creates further cost benefits, such as reduction in data center space, operational managerial overhead, support and maintenance, and so on.

Application licensing – A financial investment in newer-generation servers—which have better-performing processors within a smaller total physical socket count—can reduce the cost of purchasing licenses and maintenance renewals.

Business case next steps

• Gather data to demonstrate a clear return on investment and long-term advantages in total cost of ownership (TCO).
• Plan to recoup the capital costs of modern servers in less than a year. The returns come from improved performance, reduced power and cooling costs and simplified administration.
• Balance the need to provide new services and tight budgets. More powerful and efficient servers enable new services while lowering overall TCO cost.
• Look at overall capital and operational cost over the expected lifespan of the new servers. Don’t focus on acquisition cost. This will help to justify the decommissioning of the older equipment.
• Plan minimal disruption to the production environment. Decommission old servers with a phased approach.
• Target process and systems with the greatest immediate benefit and payback. Identify applications that need expanded capacity or improved performance and target those servers for modernization.

By Randall Lundin

Randall Lundin manages the high-end four- and eight-socket server systems in the Lenovo Enterprise Business Group. Randall came to Lenovo as part of the IBM System x acquisition where he managed IBM’s four- and eight-socket server systems. Prior to IBM, he held Product Management positions over 15 years in the software and telecom industries.