concurrency 2. Cancellation 3. An extended example3 GOALS FOR THE EXECUTORS PROPOSAL The vision: “An asynchronous analog of the STL” • A full suite of standard async algorithms based on real-world requirements: Libunifex: https://github.com/facebookexperimental/libunifex54 SUMMARY FROM PART 1 1. Vision: “An asynchronous analogue of the STL” 2. Some simple examples, intro to senders 3. The lifecycle of an async • Only algorithms that introduce concurrency need to handle cancellation. • Cancellation is asynchronous and cooperative. • Cancellation is best-effort No guarantee an operation will stop promptly

updates to the data? • Try to figure out some networking scenario to pull in the data when needed The STL currently does not have parts requiring networking and we would really like to avoid adding that updates to the data? • Try to figure out some networking scenario to pull in the data when needed The STL currently does not have parts requiring networking and we would really like to avoid adding that

with the cert you specified, common if you want to TLS with service outside mesh apiVersion: networking.istio.io/v1alpha3 kind: DestinationRule metadata: name: reviews spec: host: reviews trafficPolicy: Service ● AUTO_PASSTHROUGH: pass through the TLS traffic purely using SNI without VS apiVersion: networking.istio.io/v1alpha3 kind: Gateway metadata: name: bookinfo-gateway spec: selector: istio:

protection duration (e.g., the user-defined fn is allowed to block or take long time) Asynchronous Reclamation • Asynchronous reclamation is invoked when the number of retired objects reaches some threshold: Michael Folly open-source library: github.com/facebook/folly under synchronization/Hazptr.h Is Asynchronous Reclamation Always Enough? delete ptr shutdown_resource_X() Hazard Pointer Synchronous Reclamation reclamation ~Foo() { use_resource_X (); } Foo* ptr = new Foo; Remove ptr Is Asynchronous Reclamation Always Enough? No Asynchronous reclamation: delete ptr Using unavailable resource X Hazard Pointer Synchronous

name: nginx image: nginx EOT Rust- VMM Demo Future Works • Extend Rust-VMM to support networking and storage • Security policy integration • Develop a new fully Rust-based runtime for edge computing

chenmingyu4@huawei.com 华为 公共开发部 嵌入式软件能力中心 Applications of Rust Runtime in Mobile Overview of asynchronous Rust #1 Rust 异步简介 Ylong async runtime #3 Ylong Runtime 并发框架 目录 Table of Contents #2 社区并发框架介绍以及与移动端的不适配性 Introduction to third party Runtime crates and their incompatibility with mobile environment Rust 异步机制 Asynchronous Rust 异步并发框架是许多大型应用、系统具备的底层能力。 区别于多线程编程模型，它带来以下优势：  任务调度颗粒度更小，充分利用线程资源  更可控的线程数  单个任务资源占用：几十 await  Waker asyn c Future Waker poll Syntax sugar wake await Rust 异步机制 Asynchronous Rust Rust 异步机制 Asynchronous Rust Waker Task Future task Queue wake Worker Future.poll() Reactor fd

requests per second with minimal overhead. Fast async/await reduces the complexity of writing asynchronous applications. Paired with Tokio's utilities and vibrant ecosystem, writing applications is a

allowed)16 Coroutines  What are coroutines used for?  They simplify implementing:  Generators  Asynchronous I/O  Lazy computations  Event driven applications17 Coroutines  C++20 contains language