The data that would be shared is actually "static" data structures built from DB queries that need to be accessed multiple times per child per request. Trying to communicate over IPC for this sort of look-up would not be efficient for my needs. And when I say "static", I mean this data is to be read at start-up and will not be changed by any of the children (or parent) during the life of all processes / threads. I'm just starting on the architecture for this, so prototyping and benchmarks will consume a good portion of my time over the coming weeks. This was mostly a solicitation for technology suggestions.
In reply to Re^2: Reliable asynchronous processing
by Codon
in thread Reliable asynchronous processing
by Codon
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |