Data center assets -- such as servers, network components and power equipment -- should be retired on a routine basis to ensure application and operational workflows continue uninterrupted and at maximum efficiency. In some cases, retiring hardware simply means decommissioning it in preparation for disposal. In the majority of cases, it means making room for newer equipment that's more reliable and technologically superior to its predecessors.
Failing to replace equipment in a timely manner can be costly and can damage an organization's reputation and future business. On the other hand, retiring equipment too soon can be a waste of perfectly good hardware, incurring unnecessary effort and expense. Knowing when to retire an IT asset can be tricky because each piece of equipment is different and requires handling on an asset-by-asset basis.
IT asset retirement vs. IT asset disposition
Retiring a data center asset refers to the process of removing it from service. It is one of the final stages in the asset management lifecycle in which equipment is decommissioned or deactivated when it can no longer effectively serve its intended purpose. This is not the same as asset disposition, a step that comes after retirement to ensure assets are safely and securely destroyed, recycled or repurposed.
Retirement removes an asset from the normal data center workflow. The equipment might not be hauled off right away, but it will no longer participate in day-to-day operations. A server, for example, might still be sitting in the rack, but it will be disconnected from the network, have all its data wiped and the power turned off. Only then will the disposition phase kick in.
Benefits of retiring an asset
Organizations that wait too long before retiring their assets are putting their applications, data and overall operations at risk. For example, an IT team might be running the same WAN router they purchased eight years earlier, leaving the company susceptible to internet security threats that did not exist when the router was manufactured. Aging equipment also runs less efficiently, requires more time to maintain, causes unplanned downtimes, and affects application performance and user productivity, all of which result in higher costs.
Retiring data center assets at an appropriate time can reduce expenses because new equipment is more reliable and efficient. IT teams do not have to spend as much time maintaining the assets, and newer hardware typically uses resources more efficiently, requiring less power and cooling while reducing their data center footprint. Newer equipment is also more secure, powerful and easier to manage, leading to fewer risks, better performing applications, higher productivity and less downtime.
Criteria for evaluating retirement status
Determining when to retire an IT asset depends on the type of asset, its current level of operation and the circumstances in which it operates. One of the most obvious signs is when equipment -- or the individual components within it -- starts to fail. Managers might be able to keep things running by replacing parts or performing routine maintenance, but it's usually clear when a piece of equipment is heading toward collapse.
An unused asset is also a good candidate for retirement. For example, many organizations have servers that sit idle but are still plugged into the network and the power grid. Not only do they eat up electricity and require cooling, but they represent a security risk, often without anyone realizing that they're still churning away.
Another indication that an asset might be ready for retirement is when it can't meet current business or operational requirements. This might mean the hardware no longer performs as it once did or does not perform well enough to support current workload demands. Or it might mean the asset cannot meet current capacity, feature, security or compliance requirements.
For example, a device might no longer receive regular firmware updates that ensure its security, or it might be inadequate to support recently implemented workloads. Another possibility is that the equipment takes up too much space or uses too much energy. Such considerations are particularly acute for organizations undergoing rapid growth.
IT assets typically come with manufacturer warranties that vary in length, depending on the type and quality of the equipment. After a warranty expires, hardware is more likely to cause problems, and customers are on their own to address those issues. They might be able to purchase extended warranties, but those typically come with hefty price tags.
Despite warranty expirations, some organizations will try to make do with their existing assets rather than face the capital expenditures that come with upgrading to newer equipment. The problem with this strategy is that supporting older equipment can cost more over the long term if you take into account administrative time, loss of user productivity and higher energy consumption. At the same time, customer-facing applications might start running slower, which can translate directly to lost revenue. Worse still, if older equipment causes additional security risks for an organization, the costs in fines, penalties, lawsuits and loss of business can be astronomical.
Another consideration is whether the product and its components are still being manufactured. If they are, you might be able to find the parts you need should it be necessary to replace a component. However, once the equipment stops being made, you might have difficulty finding replacement parts, those parts might be extremely over-priced or you might have to contend with third-party components that are substandard or used.
Retiring server, network and power assets
When evaluating whether data center hardware is ready for retirement, IT teams must consider each asset on an individual basis. For example, it's generally recommended that servers be replaced within three to five years. The length of the warranty will generally determine how long an organization can expect reliable server operations. For each year beyond the warranty, the server's support costs go up exponentially. The server might also be more susceptible to security risks, especially if it's no longer receiving firmware updates.
At the same time, managers should track metrics -- such as a server's uptime and processing speeds -- to check for degrading performance. A server should also be evaluated for its ability to meet current software and workload requirements. For example, an organization might want to upgrade to Windows Server 2019 or run data-intensive analytics, but the server might not be able to support the OS or sophisticated applications. In addition, managers should evaluate whether the server is experiencing repeated problems with existing software. Even if they seem minor, the time spent resolving them can quickly add up.
The expected lifespan of network components also tends to be three to five years. Although many network components can run substantially longer, security and compliance remain top concerns and older devices tend to present greater risks. As with servers, the length of the warranty is usually the best guide on a device's recommended lifespan. In addition, once the device is no longer supported, it stops receiving software and firmware updates.
Other issues can also be a factor. For example, a router could start experiencing frequent connection problems. Restoring factory settings might help initially, but this approach generally turns out to be a temporary measure and grows increasingly ineffective the more often it needs to be done. In some cases, IT teams will need to prioritize which network components get updated first, basing their decisions on security and how much the network depends on a device.
The assets that make up the power infrastructure include a wide range of components with life expectancies that can vary significantly. For example, a power distribution unit might come with a one-year, two-year, five-year or even lifetime warranty, while an uninterruptible power supply (UPS) tends to be limited to between one and three years despite manufacturer promises for up to 15 years of operations.
Each component in the power infrastructure should be evaluated for retirement based on its unique characteristics. For example, a UPS includes batteries, capacitors, fans and a control logic board that can be failure points, while the UPS itself remains a viable system. Replacing these components eliminates the need to retire the entire UPS system. At the same time, the costs of those components can add up as the system ages, or they might become unavailable if the manufacturer stops making them, in which case, retirement is inevitable.
Moving toward retirement
An organization that fails to retire its IT assets in a timely manner risks poor application performance, compromised security, lower productivity and higher costs. The key is to find the right balance between premature retirement and putting application and operational workflows at risk. An effective retirement strategy means assessing each component individually, based on type, circumstances and recommended lifespan. When it comes to asset retirement, there is no one-size-fits-all formula, just a set of guidelines to set you on the right path.