News: P# has evolved into Coyote. This repository is now deprecated and development has moved here. Check it out!
P# is a framework for rapid development of reliable asynchronous software. The P# project, which started as a collaboration between Microsoft Research and Imperial College London, is used by several teams in Azure to design, implement and automatically test production distributed systems and services.
The P# framework provides:
- An actor-based programming model for building event-driven asynchronous applications. The unit of concurrency in P# is an asynchronous communicating state-machine, which is basically an actor that can create new machines, send and receive events, and transition to different states. Using P# machines, you can express your design and code at a higher level that is a natural fit for many cloud services.
- An efficient, lightweight runtime that is build on top of the Task Parallel Library (TPL). This runtime can be used to deploy a P# program in production. The P# runtime is very flexible and can work with any communication and storage layer.
- The capability to easily write safety and liveness specifications (similar to TLA+) programmatically in C#.
- A systematic testing engine that can control the P# program schedule, as well as all declared sources of nondeterminism (e.g. failures and timeouts), and systematically explore the actual executable code to discover bugs (e.g. crashes or specification violations). If a bug is found, the P# testing engine will report a deterministic reproducible trace that can be replayed using the Visual Studio Debugger.
Read the P# programming guide and then read about various features and topics here.
How to build
Follow the instructions to build P# from source, or just install our latest P# NuGet package.
How to contribute
We welcome contributions! However, before you start contributing, please read carefully the development guidelines.
If you are interested in using P# in your project, or have any P# related questions, please send us an email or open a new issue.