Old PCs: When it's time to die
End of life
Workshop Often, PCs are like old soldiers: they never seem to die. In many cases the old workhorses keep on going, so long as they are not touched and nothing major explodes or dies.
We know this from your comments and we see this in our research too. The upshot is that the lifespan of mainstream PCs, for users who are not too demanding, is five, six and even seven years.
By contrast, performance PCs are typically refreshed every three or four years. Even then they are often re-used as part of the pool of mainstream PCs.
It may seem good sense to keep PCs for as long as they are usable. But this strategy may be penny wise and pound foolish when taking into account the bigger picture of lost productivity, support costs, power consumption and security risks. How do we decide when enough is enough, and send old PCs to the Knackers' Yard?
One may assume that it is cheap to keep on running an old working PC. But time never stands still - the business evolves, demands change and newer technologies emerge. Meanwhile, the PC ages and slows down as it is loaded down with new or updated software.
Let's explore some of the issues in setting a PC refresh policy. We'll kick off with security - one of the most important, and often most overlooked, factors.
Many older PCs are running much the same software that they were initially configured with. Windows XP has been around since the turn of the century, but the age of some machines means they may run Windows 2000 or even variants of Windows 98 or NT. These lack many of the security improvements incorporated into modern operating systems and applications.
Ageing PCs may also affect user satisfaction and productivity. This is often a tricky issue, as salaries are not paid by the IT department.
Employees often try to “work” around performance difficulties by doing other things while waiting, such as grabbing a cigarette during a lengthy boot-up or a coffee when an application goes into hour-glass mode.
Sometimes they may even do something useful: not all workers use their PC as the central element of their job, but for those who do, the performance of their PC can have a non-trivial impact on performance and efficiency. Always factor this into discussions with the business on budget and refresh plans.
Cost of support
New PCs are not particularly expensive to buy. Much of the cost of replacement is in the provisioning and roll-out, application licences and end-user training for new applications.
As PCs move out of maintenance, any issue that needs support or attention from IT will cost money, particularly if the support means downtime for the user or desk visits from support.
Issues can often mean more in cost terms than issuing a user with a new PC when a problem arises: it may be more effective to issue a user who encounters problems with a 'new' PC rather than attempting to fix what has gone wrong.
This is even more likely to be the case as desktop virtualisation solutions become more widely used across the business.
Understanding the issues of PC lifecycle management is one thing. But few companies have a good handle on their PC, their specs, who is using them, what is running on them, and how much they cost in operations and support.
This information is complex to obtain and difficult to maintain, and we are told time and again that many businesses lack the tools needed to manage PC assets, software distribution and licensing.
Patchy insight can make it very difficult to justify a planned obsolescence and refresh strategy. The end result is that rules of thumb and gut instinct, rather than the real picture, determine the course of action.
Managing old PCs so that they fade away rather than soldier on is best accomplished by investing in modern and automated management tools, so that the real trouble spots are identified and dealt with before they have too great an effect on the users.
Today's Western brats are too picky, only fast, free PC's needed
In a Canadian city there was a drive to get corporations to donate computers to a charity whose volunteer members would recondition them and install free Linux OS and operating suite and they would BE GIVEN AWAY to underprivileged kids whose parents had little money to spend on computers.
After a while there was a considerable stack of unwanted, fully working computers. A friend, who was one of the refurbishing volunteers, told me the computers 'weren't acceptable' to these ungrateful brats 'because they were too slow for games.'
Junking them was considered, so I said rather than waste everyone's energy how about donating them to overseas students. Now there is a 40 foot container standing in a yard, which when when filled with working 3-4 year old computers, is transported here to VietNam where they are distributed to the more remote areas of this country so children can gain knowledge similar to those resident in the cities.
We have now placed over 3,100 computers in small, remote villages and recently a national drive has enabled many of them to be connected to the InterNet at no cost by a national communications company here.
We are now looking at Laos and Cambodia.
OS slow down
Many people swap their computers due to the OS getting into a mess and being over burdened with systray icons and other "helper" applications that are actually swallowing up lots of RAM.
There's so much software out there. Software that is so bloated and slow that they install so called "fast start" programs that are run upon bootup, I'm not going to name companies (hint: a PDF reader and Java). These pre-load part of the application into memory so it appears to start much faster, but just wasted RAM.
Often what is needed is a re-install of Windows and a bit more RAM in the computer.
Ah yes, the old treadmill approach
It seems to me that this article is really an apeal to refill the coffers of Microsoft. Just because the PC in question does not have a multi-core CPU or 4 Gb of memery is no reason to scrap it. Intelligent use of other O/Ss will lengthen the lifetime of said PC.
Besides aren't we all supposed to be saving the planet? If we can get more use out of existing machines why keep emitting all the noxious pollutants involved in the manufacture of a PC? As I write this an image referring to the article about the niobium/tantalum war in the Congo is sitting to its right. Surely cutting down on the use of such scarce and contentious resources is a good idea. I think that the author has been listening to the bean counters a little to much.
Back in 1987 I ordered a Mission 486DX for CAD work. It worked well and was very reliable whilst the cheapo's in accounts bought Amstrad's and wished they'd hadn't. In 1991 it was retired from CAD work, and became a word processing station, then further retired in 1993 to become part of a production line test set up. It was still chugging merrily away in 1995 when I left the company. Keep them cool, keep them clean, treat them right and they'll last years. How about security? It was never connected to a network, all data passed by 3.5" FD.
methinks the need to upgrade these days is not as prevalent as it was. The step change from 486 to Pentium was huge, IMO we don't see that these day, only small incremental changes.
Enough is enough
Enough is enough when Linux becomes too fat for the platform. Every PC should enjoy a decent OS before it dies.