Sign up now to get free exclusive access to reports, research and invitation only events.
Nominations closing in
Your computer's next point of failure might be further up the stack than you think
The early computers of the vacuum-tube age were marvels of engineering for their time. Today, we can simultaneously appreciate the advances these computers represented, making previously unthinkable computational work possible, and chuckle at what to us seems like their ludicrous size and painfully slow processing speeds.
But while you probably know ENIAC took up a whole room, you probably don't know that it just up and stopped working -- a lot. The vacuum tubes that made up its guts would burn hot and eventually fail -- and there were so many of them that this would happen every couple of hours. The reliability that solid-state electronics provided was at least as important an advance as improved speed and size.
That sinking feeling
But while we don't need to replace components in our computers ten times a day anymore, that doesn't mean there aren't points of failure to contend with. IEEE expert Tom Coughlin discusses what's called the "bathtub curve" of component failure rates: there's what's known as "infant mortality" as the bum units of a particular component die young (often in the factory itself during testing), then a long, steady stretch of few or no failures, followed by a spike as components reach the end of their design life. This allows for an expected period of high reliability -- but also assumes eventual failure. And many believe that increasingly complex electronics systems defy such neat analysis. What danger lurks among the components on your desk or in your data center?
A lot of people casually refer to any power strip as a "surge suppressor," but not all of them actually protect your equipment from voltage spikes. And even the ones that do tend to sacrifice themselves to save your computer -- often without you knowing it. "Their purpose is to protect the devices plugged into them from surges," explains Coughlin. "But that doesn't mean they will continue to do that with multiple surges. After the surge suppression function fails, they would act like a regular power strip." Coughlin notes that some -- but not all -- surge suppressors indicate whether this key function is still working.
Anything with moving mechanical parts is subject to eventual failure, and the most important thing to most people's electronic lives that fits that description is a hard drive. "If the hard drive begins to make noise, it is only a matter of time," says Danny Brugman of IgnitingBusiness.com, a Web design, technology and marketing company. "And even if it's making noises which serve as a sign of aging, the drive can last for a quite a while before it completely fails with little to no additional red flags." Nick Teeple of sever support provider Skyline Servers puts it even more grimly: "Most consumers and even small business owners don't realize that the one piece of their infrastructure that stores all of their data will completely fail in a matter of years."
But also flash drives
Don't think the industry's widespread move away from magnetic hard drives to flash memory represents a cure, though. "Even solid state drives wear down faster than most people think, though they are a lot less prone to malfunction than HDD," says Brugman. Among other issues, the actual layer of metal oxide in the solid-state drive can degrade from constant use, leading to eventual failures even though there are no moving parts rubbing up against each other. These problems can be staved off with intelligent SSD controllers, but not forever. Brugman quotes one of his college professors on the ultimate takeaway on this: "If data doesn't exist in two places, it doesn't exist."
Like the vacuum tube before it, the rechargeable lithium-ion battery has given rise to a computing revolution, making it possible to untether computers from power outlets and even fit them in your pocket. Also like the vacuum tube, they are often the first point of failure of whatever gadget they're in, thanks to the reality that you simply can only force ions from the battery's cathode to its anode so many times. Ultimately, says Andrew Bernstein, a project manager with the mobile app and marketing specialists at the Demski Group, the most you can expect from a typical phone battery is about 1,000 charge/discharge cycles -- fewer for those who use fast chargers or live in extreme climates. These limitations make research on better batteries potentially very lucrative.
Uninterruptible power supplies
Batteries can be trouble -- and what's an uninterruptible power supply but just a battery sitting aside waiting to power your equipment if the grid goes down? A UPS's battery "slowly degrades over time, but most people responsible for them never check them," says Brian McNamara of Information Business System, an IT group focused on technology solutions. "Once they are in place, they are checked off the list and forgotten about, like smoke detectors in homes. When a power outage hits, people find out the hard way ... this can lead to data loss, database corruption, and operating system failures." McNamara says that UPS batteries (like your smoke detectors) should be tested quarterly.
The functionality we expect out of our devices has slowly moved up the stack, from physical components to software: and though software doesn't "wear out" the way that the platter of a hard drive can, its usefulness can decline in a way very similar to a physical component. Joe Kelly, founder and CEO of Legal Workspace, a provider of cloud-based solutions for law firms, points to firewalls as an important case. "Many consumers or companies may get lax around firewall upkeep, thinking that once they have deployed it, they are covered. But if they are not regularly patched, then vulnerabilities will develop." The physical hardware itself may be humming along nicely, but "the firewall software is the main source of decay."
Decay in the cloud?
And let's get even further up the stack: what if the decay that will render your gadget useless isn't happening on your gadget at all? IEEE member Stuart Lipoff urges you to think about the all those apps you've downloaded to your phone and built into your lifestyle. If they "rely upon cloud services from emerging companies," he says, "they can become inoperative because the services provided by the startup firm goes out of business or they stop supporting your app." The dark side of the ease that distributed computing brings to our lives is that we rely on services run by organizations completely outside our control. (Anyone who wept bitter tears at the death of Google Reader knows what I'm talking about.)
Reduce, reuse, recycle
All of the component failures we've discussed here are just that -- failures of components. Often even when one thing goes wrong, there's plenty of useful tech left inside that gadget that seems like junk to you now. "Just because a device may no longer be in working condition, doesn't mean it holds no value," says Sean Magann, vice president of Sims Recycling Solutions. "Once removed from the workspace, IT components can either be refurbished or assessed and remarketed. Identifying this value is best done by a professional with the expertise and familiarity of the secondhand market, ensuring you're receiving the best value for all devices." Oh, and you'll be keeping working components out of landfills, too.