The final section of this chapter covers the concept of worker pools. In the preceding section on
Parallel.js, we ran up against an issue where workers were frequently created and terminated. This is a lot of overhead. If we know the level of concurrency we're capable of operating at, then why not allocate a statically-sized pool of workers that can take on work?
The first design task for creating a worker pool is to allocate the workers. The next step is to schedule the jobs as they come in by distributing them to available workers in the pool. Lastly, we'll need to account for busy states when all the workers are busy. Let's do this.
Before we think about allocating pools of worker threads, we need to look at the overarching ...