Let's now discuss the affinity between compute and storage. There is an interesting relationship between speed, capacity, persistence, and cost. When your data lives near your processor, you can start working on it immediately, as opposed to fetching it over the network. That's the promise of local storage.
There are two primary ways to store your data locally: in memory and on local drives. However, there are nuances; memory is the fastest, SSD drives are about 4 times slower than memory, and spinning disks are roughly 20 times slower than SSD drives (https://gist.github.com/jboner/2841832).
Let's consider both of these following options:
- Storing your data in memory
- Storing your data on a local SSD