Starina (named after stellina), is a general-purpose, microkernel-based, modern operating system designed for developers. It aims to be a production-ready OS, and more importantly, a fun and easy-to-understand OS where you can enjoy the development as if you are writing a Web application.
The ultimate goal of this project is to create a production-ready OS to be a good alternative to real-world OSes. To make this happen, Starina values the following principles:
- Userspace-first approach: Make OS development approachable and fun for everyone. Prioritize developer experience in the userspace, where the most OS components reside. The microkernel is just a runtime for applications.
- Simplicity over perfection: Emphasize a straightforward design which covers the most common use cases. Make it work first. Make it better later.
- Incrementally adoptable: Facilitate easy adoption of Starina by providing a seamless integration with existing systems.
This year we're focusing on cloud computing, with starina.dev running on Starina's Linux compatibility layer:
- Microkernel prototype in Rust (starina.dev running on Linux/QEMU!)
- Complete redesign and rewrite
- Rust-based zero-cost isolation (unikernel style)
- Device tree support
- Asynchronous message passing + epoll-like event driven API
- Declarative OS service discovery
- TCP/IP networking
- Virtio-net device driver
- WSL2-like Linux compatibility layer
- Linux container image support (
docker run-like experience) - Shell (in an unopinionated headless Web-based approach)
- File system server
- TypeScript (WebAssembly-based) or Swift (Embedded Swift) API
- Traditional usermode-based isolation
Install dependencies:
./setup.shBuild and run:
make runDebug with GDB (keep make run running in another terminal):
make debugStarina provides a lightweight VM based Linux compatibility library with an ergonomic std::process::Command-like API. Learn more in this blog post.
Also, make run starts a single Linux container automatically to demonstrate the Linux integration. Set LINUXRUN_IMAGE and LINUXRUN_ENTRYPOINT to specify the image name and entrypoint:
$ export LINUXRUN_IMAGE="docker://hello-world:latest"
$ export LINUXRUN_ENTRYPOINT="/hello"
$ make run
[kernel ] INFO Booting Starina...
...
[linuxrun ] INFO
Hello from Docker!
This message shows that your installation appears to be working correctly.
To generate this message, Docker took the following steps:
1. The Docker client contacted the Docker daemon.
2. The Docker daemon pulled the "hello-world" image from the Docker Hub.
(riscv64)
3. The Docker daemon created a new container from that image which runs the
executable that produces the output you are currently reading.
4. The Docker daemon streamed that output to the Docker client, which sent it
to your terminal.
To try something more ambitious, you can run an Ubuntu container with:
$ docker run -it ubuntu bash
Share images, automate workflows, and more with a free Docker ID:
https://hub.docker.com/
For more examples and ideas, visit:
https://docs.docker.com/get-started/
We (and perhaps you too) love to debate the best text editor and programming language, sometimes very seriously and passionately.
Starina is entirely written in Rust because it is "C++ with seatbelts", which is suitable for building a robust yet high-performance OS. Seatbelts are sometimes annoying indeed, but we know it saved us from countless bugs by enforcing good practices. Notably, I don't need address sanitizer when writing Rust. That's a huge factor for me.
That said, it's crystal clear that Rust (or any other language) is not the best language for everything. That's why Starina is designed to be language-agnostic, and I plan to add seamless support for other languages such as TypeScript. What if you can prototype OS components such as device drivers, as if you are writing a Web app? Isn't that cool?