“Planned obsolescence is another word for progress,” said James Jeffrey Roche. And because progress is rapid and continuous in information technology, so too is obsolescence. The promise of better, faster technology is just one reason why you need to evaluate whether it’s time to replace them on a regular basis.
Also, manufacturers design servers with a duty cycle that averages around 42 months. Actual life spans, however, can vary. Just as blue collar workers who do hard physical work have shorter life expectancies than their white collar cousins, servers that have more demands placed on them may not make it to the 42-month mark.
Finally, the marketing wizards who price new equipment and maintenance contracts do so in a way that provides an incentive for you to replace rather than keep technology that’s more than a few years old.
Given these forces that drive the need to replace servers, what’s the best way to determine whether your servers are past their useful life in your organization?
Take the Pulse of Your Server Inventory
Stretching your servers’ lives adds to your risk of slowdowns and outages. On the other hand, despite the upfront investment, replacing them with newer technology that lowers operating costs can have a positive impact on your ROI.
Thus, if possible, you need to schedule an annual server health check to determine the risks and rewards of keeping versus retiring servers. Minimally, you should conduct a server review when your warranty, maintenance contract or lease is expiring.
Here are four steps you need to take for this review.
Monitor Your Servers
To do a thorough review of your servers’ condition, you need to have been monitoring all your servers 24/7/365 over the course of a year or so. If you have a patchwork of server monitoring tools, each covering a single brand (IBM, Solaris, Linux, and more), piecing together all the information to see the big picture will be more challenging. On the other hand, a single all-knowing monitoring solution makes it easy to find all your servers and their vital information rapidly — model numbers, types of technology, capacity, memory and peak usage. Also, with historical data, you can assess trends that help you to predict future needs.
Locate Servers Over Three Years Old
Use your data to find servers that are three or more years old. You may discover, for example, that you have some IBM POWER5 and POWER6 servers, which IBM stopped supporting several years ago. To keep them in service will require some hefty maintenance costs.
Look for Poor Performers
Look at the performance of all your servers. Remember, some of the harder workers don’t last as long. Are any of them crashing or causing slowdowns? If so, that could lead to poor service for associates or customers.
Create Virtual Groups and Determine Your Needs
You now have your candidates for replacement, the older servers and poor performers. Use your server monitoring software to tag and virtually group them.
Find out how much capacity they’re using. Servers have a tendency to multiply, so you are likely to find wasted capacity. Even with yesterday’s technology, it is possible you do not need as many servers as you have. And now that we’ve moved along the technology curve, faster speeds mean you’ll need even fewer servers.
So run a simulation to determine how many servers you need based on your capacity needs and today’s technology’s performance.
Keep or Replace?
Weigh the pros and cons of staying with your current server inventory or refreshing it.
Let’s say, for example, you discover you have 100 servers that are more than three years old. Ten of them are out of capacity, so you need to move them to larger, faster environments. But what about the rest of them?
- The Costs of Keeping Servers
Here’s one consideration that looms large — they have all rolled off their warranty. To ensure that they all stay up and running, you need to buy a $500,000 a year maintenance contract. And that’s just the beginning. It will probably increase to around $600,000 or $700,000 in future years. Bottom line? The operating costs of aging servers are high.
- The Rewards of Replacing Servers
How do the costs of keeping older servers compare to those for buying new servers?
To find out, play with some scenarios in your server monitoring software. You may discover you only need 25 new servers to take care of the work of the 100 servers under review plus give you enough capacity for the next three years. The price of 25 servers, the associated cores, network adapters and disk adapters tallies up to $750,000. That cost includes a three-year warranty, eliminating the need for a maintenance contract.
Since it’s only 50 percent more than a one-year maintenance contract, buying new servers is already starting to look like a winner.
Also, 25 servers will take up less floor and rack space in your data center. So if you are renting space, or can downsize a facility your company owns, you will realize savings. You’ll also cut your electric bill and consume less of your operations and systems employees’ time for administration.
There are a couple more rewards. New servers significantly reduce your chances of suffering costly slowdowns and outages. In addition, the latest technology may be slightly faster than your current servers. Perhaps it’s just a minor improvement in speed that adds five minutes of productivity a day to each employee. If, however, you have 1000 associates, that’s 5000 minutes. And 5000 minutes is equivalent to about ten new full-time employees.
Given this rosy upgrade scenario, would there be any reason not to take the leap? Yes. A server migration has the potential to disrupt operations, so your IT staff need to have time to plan, implement and monitor it. If they are up to their ears in other projects, you might want to minimize the migration to those servers at risk of causing downtime and slowdowns. Pay for the maintenance plan for the others and keep them running for another year or so until you have time to deal with a server migration.
Technological advancement, short server life-cycles and market pricing that favors replacement, all make it necessary to review your server inventory regularly and determine whether to replace some of them. Because of this reality, you need to monitor your servers around the clock with a holistic tool, then weigh the costs and benefits of buying new technology versus those of maintaining your existing inventory. By doing so, you can decrease costs and increase performance.