Is it time for a data center hardware refresh?

Options are plentiful when you're considering whether to refresh older data center hardware, even when IT budgets are not.

Data center hardware ages over time and will eventually need to be replaced, but the decision to replace older hardware has become more challenging in recent years. In theory, it’s really a simple equation. If the hardware is more than N years old, it’s time to replace. But given the state of business and finances in most organizations, many are seeing an end to the traditional three- or four-year cycle of replacing all their hardware simply because of budget constraints. Holding off on that next upgrade means you may be able to just continue paying maintenance -- and save large capital outlays for new machines.

Still, it’s impossible to stick with the same hardware indefinitely. Simply continuing “as is” amid rapidly evolving and changing technology just doesn’t work. New applications demand the performance and features often found in the newest systems. Similarly, the rapid adoption of technologies like server virtualization make it increasingly difficult to stick with traditional hardware provisioning tactics (e.g., “one app, one machine”), especially when you’re required to reduce costs as well.

Evaluating the need for a technology refresh
From a practical standpoint, all administrators will find themselves squeezed between the same three issues: increasing business requests and requirements (more computing); requirements to reduce IT spending (lower budgets); and demands for more service speed and efficiency. There’s always a temptation to make cost reduction the priority, usually with pressure from upper management. However, one of the results of increased efficiency is reduced costs, so it’s the perpetual struggle to make strategic investments (such as technology refreshes) that can ultimately save money for the organization. Implementing chargeback can significantly reduce or defray the direct spending by IT as well.

Given these factors, prioritizing comes down to one simple concept: what can you implement with the budget, staff, expertise and infrastructure that you have? You may not have the staffing to design and deploy a chargeback system, or you may not have the financial resources to do a major rebuild of your data center. But at the same time, you can’t be afraid to reach as well. Some items might cost more to implement in the short term, but actually save money over the course of the technology cycle. And in many organizations, it’s as much about shifting costs away from IT as it is reducing them.

When the technology refresh can’t wait
Things can get sticky when new hardware is an actual requirement. Many organizations face aging infrastructures and equipment they can no longer buy service contracts for. Third-party maintenance might keep older equipment online longer, but that’s not your only option when it comes to a technology refresh.

Consider leasing. For example, organizations might use leasing agreements to make obtaining enterprise systems affordable -- usually on a 36-month cycle. Depending on your equipment, purchasing all new hardware on a 48- or even 50-month lease cycle may make more sense. Almost every manufacturer has very recently done complete refreshes of their product lines, making it the perfect time for a long-term lease. Most of these systems are just entering their lifecycles or are only a year or two into it, making it unlikely they’ll be subject to increased support costs within the span of the lease.

Look for creative financing. More and more leasing companies are willing to get creative with financing options, extending well past 36 months, and offering discounts for larger buys or combinations. There’s also stiff competition between vendors and manufacturers, and almost everyone will provide significant competitive upgrade credits in one fashion or another. Don’t be shy about asking about them, or what offers might be available closer to the time you’ll actually be writing out purchase orders. If they know that you’re having trouble making it work, they might try to help bridge the gap.

Peek at the secondary market. It may seem macabre to some, but with the recession, numerous “like-new” systems were put into the secondary market when companies went bankrupt or simply sold off systems and parts they no longer needed or no longer could afford when the downturn affected their business. These lightly used systems are often taking up valuable warehouse space at leasing companies, who are keen to unload them. Systems sitting in warehouses aren’t generating revenue for anyone, and are often available for a fraction of the cost of new, and still qualify directly for service contracts without recertification.

Putting off the technology refresh
For some organizations, buying new hardware simply isn’t an option, or may even be a bad option. Maybe the budget just isn’t there, there isn’t support for it outside of IT or any number of reasons. There are still options available when the situation forces you to minimize costs and make the most of what you have.

Consider upgrades for existing hardware. Even if you can’t buy direct from the original manufacturer, many third parties can provide new or recertified components for upgrading those systems. And for businesses that have seen a drop-off in sales or lower revenue projections, it may actually make more sense to ride things out by simply upgrading the existing hardware rather than committing to new hardware purchases.

Think about virtualization and commodity servers. Sizing a technology refresh is a complicated task, requiring a thorough understanding of what the business plan means to IT requirements. Many organizations still aren’t certain; they know where they’d like to be -- and where they should be -- but not when and at what rate they’ll be there. For example, buying the right amount of power depends on knowing how much power is really needed. Uncertainty about business goals and needs can prompt even the most frugal organization to overbuy a technology refresh by hundreds of thousands of dollars, or rush buying and deployments when the sizing was done against the current trough rather than the expected peak.

When that’s the case, there are still options to improve your utilization and get more out of aging hardware. Some workloads may be candidates for being migrated to lower-cost x86 servers, freeing up more expensive computing resources where they’re needed. Shifting workloads can be increasingly critical as application requirements continue to grow. Server virtualization is also easy and inexpensive to add to an environment, making it possible to improve efficiency by increasing the work done per watt of power and the number of applications running on each system. Windows environments can use Hyper-V, Citrix and VMware virtualization, and every Unix system currently on the market now has multiple virtualization technologies available.

Think outside the box
There are usually numerous options that can fit new hardware into the data center or squeeze a bit more life from existing hardware. And when these resources seem to be fully exhausted, it may be possible to incorporate new systems in staggered fashion, with all the leases ending at the same time -- yes, the same time. So you might just replace  the ERP server and upgrade its associated storage this year, then move on to other systems over the next two years, shifting workloads as appropriate. If all the leases come up at the same time, you can then (years down the road) opt to renew the leases, buy new hardware or split the difference as your situation allows.

Dig Deeper on IT infrastructure management and planning