|Doc [Bastion Executor]||Discord|
|Doc [LightProc]||Build Status|
Bastion is a highly-available, fault-tolerant runtime system with dynamic, dispatch-oriented, lightweight process model. It supplies actor-model-like concurrency with a lightweight process implementation and utilizes all of the system resources efficiently guaranteeing of at-most-once message delivery.
Bastion comes with a default one-for-one strategy root supervisor. You can use this to launch automatically supervised tasks.
Include bastion to your project with:
bastion = "0.4"
Official documentation is hosted on docs.rs.
Check the getting started example in
Examples cover possible use cases of the crate.
io_uringsupport and much more...
futuresand lightproc layers.
If you answer any of the questions below with yes, then Bastion is just for you:
Bastion Ecosystem is here to provide you a way to customize it. If you don't need to, or if you are a newcomers, you can install and use Bastion without knowing how everything works under the hood. We hope to find you on this section soon.
Nuclei is proactive IO system that can be independently used without executor restriction. It is also powering Bastion's IO system. You can learn more about Nuclei here, check out Nuclei's repo for more sophisticated use cases.
LightProc is Lightweight Process abstraction for Rust.
It uses futures with lifecycle callbacks to implement Erlang like processes and contains basic pid to identify processes. All panics inside futures are propagated to upper layers.
Bastion Executor is NUMA-aware SMP based Fault-tolerant Executor, highly-available and async communication oriented.
It's independent of it's framework implementation. It uses lightproc to encapsulate and provide fault-tolerance to your future based workloads. You can use your futures with lightproc to run your workloads on Bastion Executor without the need to have framework.
Agnostik is a layer between your application and the executor for your async stuff. It lets you switch the executors smooth and easy without having to change your applications code. Valid features are
runtime_nostd (coming soon).
Runtime is structured by the user. Only root supervision comes in batteries-included fashion. Worker code, worker group redundancy, supervisors and their supervision strategies are defined by the user.
Supervision strategies define how child actor failures are handled, how often a child can fail, and how long to wait before a child actor is recreated. As the name suggests, One-For-One strategy means the supervision strategy is applied only to the failed child. All-For-One strategy means that the supervision strategy is applied to all the actor siblings as well. One-for-one supervision is used at the root supervisor, while child groups may have different strategies like rest-for-one or one-for-all.
We use Discord for development discussions. Also please don't hesitate to open issues on GitHub ask for features, report bugs, comment on design and more! More interaction and more ideas are better!
All contributions, bug reports, bug fixes, documentation improvements, enhancements and ideas are welcome.
Licensed under either of
at your option.