It's nice to see that Spark is on par with Unix tools on low data volumes. The Spark program can be applied as-is for data volume 10 times or 100 times bigger

It's nice to see that Spark is on par with Unix tools on low data volumes. The Spark program can be applied as-is for data volume 10 times or 100 times bigger without being modified. You would have to reinvent Spark if you started from Unix tools and applied it to bigger volume of data