A peer-to-peer cluster actor system implementation for Swift.
NOTE: This project provides a cluster runtime implementation for the distributed actors language feature.
Important: This library is currently released as beta preview software. While we anticipate very few changes, please be mindful that until a stable 1.0 version is announced, this library does not guarantee source compatibility.
We anticipate to release a number of 1.0.0-beta-n releases during the beta phase of Swift 5.7, before releasing a source stable 1.0 soon after.
Most APIs and runtime are rather stable, and have proven itself for a long time already. Most of the remaining work is around making sure all APIs work nicely with the latest revision of the distributed actor language feature.
Important: Please ignore and do not use any functionality that is prefixed with an
_(such as types and methods), as they are intended to be removed as we reach the stable 1.0 release.
Distributed actors are an extension of the "local only" actor model offered by Swift with its actor keyword.
Distributed actors are declared using the distributed actor keywords (and importing the Distributed module),
and enable the declaring of distributed func methods inside such actor. Such methods may then be invoked remotely,
from other peers in a distributed actor system.
The distributed actor language feature does not include any specific runtime, and only defines the language and semantic rules surrounding distributed actors. This library provides a feature-rich clustering server-side focused implementation of such runtime (i.e. a DistributedActorSystem implementation) for distributed actors.
To learn more about both the language feature and library, please refer to the reference documentation of this project.
The primary purpose of open sourcing this library early is proving the ability to implement a feature complete, compelling clustering solution using the distributed actor language feature, and co-evolving the two in tandem.
You can refer to the Samples/ directory to view a number of more realistic sample apps which showcase how distributed actors can be used in a cluster.
The most "classical" example of distributed actors is the SampleDiningPhilosophers.
You can run it all in a single node (run --package-path Samples/ SampleDiningPhilosophers), or in 3 cluster nodes hosted on the same physical machine: run --package-path Samples/ SampleDiningPhilosophers distributed. Notice how one does not need to change implementation of the distributed actors to run them in either "local" or "distributed" mode.
Please refer to the rendered docc reference documentation to learn about distributed actors and how to use this library and its various features.
Note: Documentation is still work in progress, please feel free to submit issues or patches about missing or unclear documentation.
This library requires beta releases of Swift (5.7+) and Xcode to function property as the distributed actor feature is part of that Swift release.
When developing on macOS, please also make sure to update to the latest beta of macOS, as some parts of the Swift runtime necessary for distributed actors to work are part of the Swift runtime library which is shipped with the OS.
Distributed actors are not back-deployed and require the latest versions of iOS, macOS, watchOS etc.
When developing on Linux systems, you can download the latest Swift 5.7 toolchains from swift.org/downloads, and use it to try out or run the library like this:
$ export TOOLCHAIN=/path/to/toolchain
$ $TOOLCHAIN/usr/bin/swift test
$ $TOOLCHAIN/usr/bin/swift run --package-path Samples/ SampleDiningPhilosophers dist
Latest (beta) Xcode releases include complete support for the distributed language syntax (distributed actor, distributed func), so please use the latest Beta Xcode available to edit the project and any projects using distributed actors.
It is possible to open and edit this project in other IDEs, however most IDEs have not yet caught up with the latest language syntax (i.e. distributed actor) and therefore may have trouble understanding the new syntax.
VSCode
You can use the Swift Server Work Group maintained VSCode plugin to edit this project from VSCode.
You can install the VSCode extension from here.
The extension uses sourcekit-lsp and thus should be able to highlight and edit distributed actor using sources just fine. If it does not, please report issues!
CLion
The project is possible to open in CLion as a SwiftPM package project, however CLion and AppCode do not yet support the new distributed syntax, so they might have issues formatting the code until this is implemented.
See also the following guides by community members about using CLion for Swift development:
The project currently is emitting many warnings about Sendable, this is expected and we are slowly working towards removing them.
Much of the project's internals use advanced synchronization patterns not recognized by sendable checks, so many of the warnings are incorrect but the compiler has no way of knowing this. We will be removing much of these internals as we move them to use the Swift actor runtime instead.
Documentation for this project is using the Doc Compiler, via the SwiftPM DocC plugin.
If you are not familiar with DocC syntax and general style, please refer to its documentation: https://developer.apple.com/documentation/docc
The project includes two helper scripts, to build and preview documentation.
To build documentation:
./scripts/docs/generate_docc.sh
And to preview and browse the documentation as a web-page, run:
./scripts/docs/preview_docc.sh
Which will result in an output similar to this:
========================================
Starting Local Preview Server
http://localhost:8000/documentation/distributecluster
Integration tests include running actual multiple nodes of a cluster and e.g. killing them off to test the recovery mechanisms of the cluster.
Requirements:
brew install coreutils to install stdbufThis project requires Swift 5.7+.


职场AI,就用扣子
AI办公助手,复杂任务高效处理。办公效率低?扣子空间AI助手支持播客生成、PPT制作、网页开发及报告写作,覆盖科研、商业、舆情等领域的专家Agent 7x24小时响应,生活工作无缝切换,提升50%效率!


多风格AI绘画神器
堆友平台由阿里巴巴设计团队创建,作为一款AI驱动的设计工具,专为设计师提供一站式增长服务。功能覆盖海量3D素材、AI绘画、实时渲染以及专业抠图,显著提升设计品质和效率。平台不仅提供工具,还是一个促进创意交流和个人发展的空间,界面友好,适合所有级别的设计师和创意工作者。


零代码AI应用开发平台
零代码AI应用开发平台,用户只需一句话简单描述需求,AI能自动生成小程序、APP或H5网页应用,无需编写代码。


免费创建高清无水印Sora视频
Vora是一个免费创建高清无水印Sora视频的AI工具


最适合小白的AI自动化工作流平台
无需编码,轻松生成可复用、可变现的AI自动化工作流

大模型驱动的Excel数据处理工具
基于大模型交互的表格处理系统,允许用户通过对话方式完成数据整理和可视化分析。系统采用机器学习算法解析用户指令,自动执行排序、公式计算和数据透视等操作,支持多种文件格式导入导出。数据处理响应速度保持在0.8秒以内,支持超过100万行数据的即时分析。