Slaying Dragons #25 – Redundancy and Self Diagnosis

John Watkinson considers redundancy and self diagnosis

One of the advantages of digital technology is that the equipment can be made self-diagnosing. It can log errors and in some cases correct them or at least continue working at reduced performance.

The efficiency of modern audio production would simply not be possible without digital technology. Our audio recordings are protected by error correction so that random errors don’t cause pops and clicks, and our file server might use RAID technology where an entire hard drive could fail without any data being lost. We exchange files over networks, which have multiple routes so they can survive single point failures.

This approach is also found in cars, although for different reasons. Cars had to become self-diagnosing because their owners neglected them. One of the first signs of neglect was increased emissions, so it was ironic to find people who couldn’t be bothered with maintenance protesting about pollution from cars.

Modern cars have a whole array of sensors to check everything is working properly, but they don’t generally die at the roadside if one of the sensors fails, because the system is smart enough to figure out that the sensor signals are implausible and to keep going, usually at reduced power. The modern airliner is a masterpiece of redundancy and self-diagnosis that contributes immensely to safety.

But features such as the above only come in to being, and are only retained, if there is sufficient impetus to do so. There are now plenty of reasons to think that the impetus is lacking in places and that redundancy, fault tolerance and reliability are not what they used to be in many areas.

Owing to rough seas that washed out a section of coastal track, we found that our rail network isn’t in fact a network…

(contd)
**To read the full article log in to the members’ site**