Answers

1. A thread has it's own memory and processes CPU tasks. Multiple threads can process multiple CPU computations at the same time. Async tasks have their own state and can be polled to see if they are completed or not. Async tasks are usually for non-blocking tasks such as waiting for a response from a network. Because these async tasks are non-blocking, a single thread can handle multiple async tasks, looping through and polling tasks to see if they are finished or not.
2. We can create a non-blocking sleep future by creating a struct that has a field with the time that the struct was created, and another field for the duration of the sleep. We then implement the Future trait for the struct where the poll function gets the current time, calculates the time elapsed from the created field, and return a Ready if the duration has passed, or a Pending if the duration has not passed.
3. A typical web service creates a TCP listener on a port in the main thread. The service then starts an...