Article looks interesting, however I have several questions:

1. If we are talking about making requests, they are usually done for fetching the data. Pool doesn’t seem to care about the order which makes it not that great for fetching data from lots of sources.
2. Performance testing looks invalid due to randomness of execution time for all the requests. Instead of pure comparison of the algorithms we check a factor of luck.
3. Why not use native private fields for internal methods definition instead of underscores? 🙃