What is the primary purpose of Spark's in$-$memory processing capability?

$1$ point

To reduce disk$-$based I$/$O costs

To improve data ingestion performance

To support complex data transformation tasks

To enable real$-$time data stream processing