Its no secret. Reality today in IT looks like this: lowering costs, reducing spending and if you must spend money, spending it on things that have measurable value. Those goals are laudable, yet one of the most common practices is one of the most shortsighted and potentially costly: delaying replacement of servers and PCs. While delay offers the benefit of reducing capital expenditures or out-of-pocket costs for system leases, there are many real, less obvious costs in dealing with older machines. These costs include soft- and hard-dollar items that arent always considered during the upgrade decision process.
The largest single area of cost for older equipment is increased support. For servers, its not unusual to find that system stability and uptime become compromised as patches, upgrades and new utilities are added. While many older servers remain stable, in many other cases, its an almost daily battle to ensure uptime. Further, rolling out new production systems can often cause downtime as incompatibilities or conflicts are raised based on the vendors inability to test for every older version of operating systems and middleware.
The cost of older PCs can be substantial. One CIO found that PCs that are 4 or more years old cost $46 per month more to support than those less than 2 years old. The additional cost is primarily due to system lockups and instabilities caused by DLL conflicts and driver incompatibilities. In addition, as Windows 98 nears the end of its life, providing support for Windows 98 systems will be more difficult and costly.
Another area where additional costs are common is the increased amount of testing needed when pilot applications are rolled out. Older systems add costs here in a couple of ways. The first is obvious. The more “system images” you have in the inventory, the more testing that needs to be done. In cases where the new software demands high-priced talent, those costs can easily eclipse the price of a few new systems. The second area of cost increase is when new applications or tools arent stable with older versions of middleware or operating systems and thereby “force” an upgrade. This adds cost and substantial complexity, as you now have to change two elements at the same time. If a problem happens, its harder to pinpoint the source.
There are, of course, other issues that range from the availability of spare parts and accessories to end-of-life support. The fact is, there are reasonable lifetimes for all hardware. Todays economic conditions have made us go past these boundaries in many cases. In trying to save money, we may be spending more. If youve ever owned an old car, you know a point comes when repair is no longer worthwhile. Look at what youre spending on old systems. You might be able to free your IT budget by junking the old Yugo and buying a new Chevrolet.
Aaron Goldberg is vice president of Ziff Davis Medias Market Experts Group.