The lifespan of the traditional desktop computer is expanding. 15 years ago the pace of change was so fast you have 5 years between deploying a desktop computer and it being so out of date you wrote it off as junk. While a computer first unboxed and placed on someone’s desk in 2009 isn’t be cutting edge today in 2014, it will probably run Windows 8.1 with Office reasonably well.
The minimum hardware requirements for Windows 8.1 aren’t all that different from the hardware requirements for Windows Vista released 7 years previously. Both require a 1 GHz processor, 1 GB of RAM, and around 16 GB of hard disk space. Perhaps a reflection of this is that average desktop computer performance over the last few years hasn’t improved at the sort of radical rate we saw in the 80’s, 90’s and early 2000’s. There are a growing band of commentators that suggest trend is likely to continue, with overall performance improvements becoming less noticeable, especially as more attention is paid to metrics around power consumption and battery life. It’s unlikely that the next version of Windows is going to have a substantially higher minimum hardware requirement than Windows 8.1 does, which probably means that a high end system from 2007 is likely to be able to adequately host Windows vNext.
Another thing that influences hardware lifespan, beyond increasing performance requirements, is increasing component reliability. Simply put, in the last decade, computer hardware has become more reliable, with longer mean time between failure (MTBF) for components. The average lifespan of a desktop computer’s hard drive today is substantially greater than the average lifespan of a desktop computer’s hard drive 15 years ago. Components are lasting longer and need replacement less frequently.
In the past, many organizations only switched OS when they retired computer hardware. One of the likely factors in XP’s longevity was the longevity of the hardware that ran the OS. The reasons for performing an in-place upgrade to a new OS on the same hardware are a lot less compelling to performing a migration to a new OS on new hardware when the old hardware is ready for the scrapheap.
It’s not unreasonable to assume that the desktop computers that are being deployed today will be running hardware that is likely to be quite capable of running the majority of workloads in 2020. If these computers are being deployed with Windows 7 today, it’s going to take something pretty special to get organizations to bother going through the process of replacing Windows 7 with another OS before it is absolutely necessary to do so.
In talking to a number of desktop administrators who still have substantial XP deployments, I’ve found that the primary driver of them moving off Windows XP hasn’t been features of the newer OS, or security concerns, but instead the primary driver of their final migration has been to meet compliance obligations.
I suspect that we’ll see something similar in 2020 when Windows 7 reaches end of extended support. That a fair percentage of organizations running the OS have stuck with it because they haven’t until that point felt compelled to move to anything newer because their hardware, perhaps even a decade old by this point, still allows people in these organizations to do their jobs.