Dear all,
While working on Software FMEAs we realized that each piece of software made certain use of common resources: Execution time, RAM, ROM, etc...
If a failure of a characteristic of that piece of software is defined as:
Its mother function will only fail in a way that it will take too long to execute, but most importantly that means that many of the other functions that were supposed to be performed in the same time lapse will simply not be executed at all. So we defined the following failure for these other functions:
So now, as we are working with APIS for example, and building a failure net in which each function taking too long results in the other functions not being able to execute, we can have for N functions, N*N failure net links and it will become a big spaghetti dish...
Would you maybe know of any other nice and more practical way to handle this?
While working on Software FMEAs we realized that each piece of software made certain use of common resources: Execution time, RAM, ROM, etc...
If a failure of a characteristic of that piece of software is defined as:
Execution time is too long
Its mother function will only fail in a way that it will take too long to execute, but most importantly that means that many of the other functions that were supposed to be performed in the same time lapse will simply not be executed at all. So we defined the following failure for these other functions:
Cannot be executed because of lack of resources
So now, as we are working with APIS for example, and building a failure net in which each function taking too long results in the other functions not being able to execute, we can have for N functions, N*N failure net links and it will become a big spaghetti dish...
Would you maybe know of any other nice and more practical way to handle this?