An insurer's pricing engine recalculates every policy's premium overnight on EC2. Monitoring shows the processors pinned at full use for the whole run while memory and disk stay mostly idle. The engine's code runs only on standard CPUs. Which instance category should the team move to?
Choose an answer to check it. Afterward, select any option to review its explanation.
Memory optimized instances help when large data sets are processed in memory. Monitoring shows memory mostly idle here, so more memory would not shorten the run.
Heavy number crunching makes accelerated computing tempting, since accelerators can speed up calculations such as floating point math. They do that through hardware co-processors, and this engine's code runs only on standard CPUs, so it would not use them.
The run is limited by processor work while memory and disk sit idle, which is a compute-bound workload. Compute optimized instances are meant for exactly that, giving the engine high-performance processors.
Storage optimized instances suit heavy read and write access to large data sets on local storage. The disks are mostly idle in this job, so faster local storage is not the bottleneck.
id q21d598cfaa