Hi… How does Vapor handle big data analytics?
(1) is it I/O intensive or does it do in-memory processing to avoid writing to a disk?
(2) does it work on a distributed system to scale on multiple cores and nodes (think PySpark)?
(3) Will Vapor benefit from GPUs? more RAM?
Answers to (2) and (3) are straightforward:
– 2) VAPOR has limited abilities to utilize multi cores, and no ability to utilize multi nodes.
– 3) VAPOR benefits most from beefy GPUs, and then the amount of RAM.
Answer to (1) is complicated. You can read Section 4 of this article, which provides detailed discussion.