I am not sure if this is the right place to ask this question. I am asking it anyway..

In Aircraft Airbus 320, control inputs from the pilot's sidestick are processed by two computers for different purposes. in case of failure of one computer, other computer can handle the total processing job perfectly.

My question is ' why the processing-job is divided in the first place'? What the advantages of such division in generic systems like this one?

  • 3
    $\begingroup$ Haven't you just answered your own question? ('in case of failure ...') $\endgroup$ – Matt L. Jun 26 '13 at 7:14
  • $\begingroup$ in case of failure, processing job could shift to another computer which is the exact replica of the first one. I mean, we can have just one computer for all processing and have its copies for back up. why have two computers, and for back up move to one? $\endgroup$ – KawaiKx Jun 26 '13 at 7:32
  • 4
    $\begingroup$ @Saurabh: because this is a control system for actually flying the airplane, not a laptop with harddrives. If the first system fails you need to switch over to the other one almost instantaneously. The state that needs to be moved is actual sensor measurements (windspeeds, pressures, voltages, mechanical angles) representing the physical state of the airplane. Backups of what the state was a few minutes (or even a few seconds) ago are irrelevant. $\endgroup$ – Wandering Logic Jun 26 '13 at 12:39
  • $\begingroup$ @everyone. The question asks about the behaviour prior to the failure. You may sleep one computer, you may assign it the same job, you may divide the job. The question why of the 3 options, the 3rd was chosen instead than first two? Which way do 'in case of failure' and 'redundancy' answer the question? How dividing the job instead of assigning the same jobs speeds up the recovery? $\endgroup$ – Val Jun 27 '13 at 12:22
  • 4
    $\begingroup$ Not about signal processing $\endgroup$ – endolith Jun 27 '13 at 18:25

The division of processing is to reduce the likelihood of failure. But exact replicas have exactly the same failure modes, thus providing less redundancy.

| improve this answer | |
  • $\begingroup$ so it means the division of processing job is not for redundancy, as it seems. since redundancy kicks in only after a failure has occur. here we are talking about the state before any failure. dividing the processing job does not put the entire load on one computer, thus reduces the likelihood of failure. $\endgroup$ – KawaiKx Jun 26 '13 at 16:34

Not the answer you're looking for? Browse other questions tagged or ask your own question.