Your organization runs a fault-tolerant batch processing application responsible for handling large volumes of data. The computations are distributed across hundreds of VMs to enable parallel processing. You aim to implement a cost-effective solution for scaling data processing. What should you do?