Async Rust

Chapter 3. Building Our Own Async Queues

Although we have explored basic async syntax and solved a problem using high-level async concepts, you still might not be completely sure of what tasks and futures really are and how they flow through the async runtime. Describing futures and tasks can be difficult to do, and they can be hard to understand. This chapter consolidates what you have learned about futures and tasks so far, and how they run through an async runtime, by walking you through building your own async queues with minimal dependencies.

This async runtime will be customizable by choosing the number of queues and the number of consuming threads that will be processing these queues. The implementation does not have to be uniform. For instance, we can have a low-priority queue with two consuming threads and a high-priority queue with five consuming threads. We will then be able to choose which queue a future is going to be processed on. We will also be able to implement task stealing, whereby consuming threads can steal tasks from other queues if their queue is empty. Finally, we will build our own macros to enable high-level use of our async runtime.

By the end of this chapter, you will be able to implement custom async queues and fully understand how futures and tasks travel through the async runtime. You will also have the skills to customize async runtimes to solve problems that are specific to you and that a standard out-of-the-box runtime environment might not be ...