Is computer doom coming in 2038?

If it goes unchecked, 2038 could cause widespread computer failure. Source: Supplied

FOR those of us who were old enough, we all remember the hysteria over the Y2K bug at the turn of the century. While the fallout wasn't as bad as certain media were contending at the time, it did have an impact on some computer operating systems around the world. Although Y2K ultimately proved be no more than a mere hiccup in our heavily computerised civilisation, there is a another programming limitation that could prove to be very disruptive if it goes unchecked.

The problem again lies in the time telling capabilities of computer servers. A vast majority of computer servers use the same system that stores the date and time in a 32-bit integer that counts the number of seconds since January 1st, 1970 — often referred to as the Epoch time.

The problem is that on March 19th 2038, at precisely 03:14:07 (coordinated universal time) the clocks will reach the largest number representable by a 32 bit integer.

So what happens then?

We don't know. As things are, it is most likely that many computers will not be able to tell the difference between the year 2038 and 1970. To continue to count the seconds the values will start to be stored in negative counting up from -2,147,483,647 to zero. As most systems will not be able to cope with this change it is likely that they will fail.

How they will fail is unknown. Some computers may continue to work just fine with an incorrect time code, while others that rely on a precise date and time could shut down altogether.

The biggest issue of concern, like the Y2K bug, is the potential for computer systems that control crucial infrastructure to stop working all at the same time.

Should we be concerned?

While the 2038 problem is definitely legitimate, with well over two decades to prepare, it is unlikely that it was cause any serious issues.

By the year 2038 many 32-bit systems will naturally wear out or be replaced in the next 23 years. Those systems that might not will need changing ahead of time.

Infrastructure is likely to be the biggest headache to fix (devices in power stations for example) but planning the change far enough in advance should remove most big problems.