Consider a pipeline in which each task gets input as a file and produces output as another file. If the programs run for each task are free of side$-$effects, how can you maximise the speed at which you can run a job through this pipeline?

Which of the following improves the speed, assuming you get the resources needed?

Write the option name?

Distribute the programs on many machines

Use a database to write the intermediate records, instead of files

Use an SSD to store the files.

Setup a in$-$memory file system and run on that file system