A startup has built an operating system that treats the whole OS as a library, and the idea is so clean it reads like science fiction. The company calls the project FTL, and it promises to make lightweight containers as secure as virtual machines while keeping everything fast. The pitch is simple: instead of wrestling with low-level code, developers can add features, fix bugs, and upgrade the kernel itself as easily as writing an app.
Whether that bet pays off depends on how well the design holds up in practice.
What FTL Actually Is
FTL is a userspace OS design. That means each container runs its own version of the OS as a shared library, implementing most of the familiar concepts: Linux processes, the file system, and TCP/IP. The kernel itself stays thin, providing only a minimal interface for system calls.
Think of it like a hypervisor that lives inside the user space rather than below it. Containers get isolated from each other with a hardware-based boundary. The company claims this isolation is better than what existing monolithic kernels offer.
Why a Library Matters
The userspace OS design has a practical payoff. Because the OS is a library, developers can extend most of Linux kernel features without digging into kernel or eBPF programming. Adding print statements, applying security updates, and adding new features all become quick and safe operations.
In FTL, the OS is just a library. That means upgrades happen at the application layer, not the kernel layer. The company describes this as unlocking new OS-level abilities in applications without sacrificing performance.
The Linux Compatibility Problem
The company says FTL is compatible with Linux binaries. That matters because Linux remains the foundation of most production workloads. A system that can’t talk to existing software is useless.
The proof point: the Rust-based HTTP server serving the company’s own website is a Linux application running on FTL. You can also run Unikernel-like specialized applications without POSIX abstractions. That split — supporting standard Linux apps alongside bare-metal-style workloads — shows the range of what the system can handle.
Security Through Separation
The company’s stated goal is to make lightweight containers as secure as VMs. That is a tall order, and the company frames the design as a combination of microkernel flexibility and monolithic kernel simplicity.
The security claim rests on the isolation model. Each container runs its own userspace OS instance, and the kernel enforces a boundary between them. The company describes the isolation as hypervisor-like.
The Tradeoffs
No design is perfect. The company is trading some of the raw simplicity of a monolithic kernel for the security and flexibility of a microkernel approach. Whether that tradeoff holds up at scale is an open question.
The performance picture is also incomplete. The company promises that lightweight containers can match VM security without sacrificing speed, but the source does not include benchmark data or real-world latency numbers. Those details matter for anyone building a production workload on top of this system.
Who Cares About This
Developers who want to move faster on feature delivery are the obvious target. The ability to upgrade the OS itself as easily as deploying a new app removes a major friction point in the development cycle.
The company’s own website is running a Rust-based HTTP server on FTL, so they are eating their own dog food. That is a small signal of confidence, but it is not proof of production readiness.
What the Source Does Not Say
The company’s own description is focused on the idea, not the full technical breakdown. The company wants people to understand the idea, not just the code. That is a reasonable strategy, but it leaves readers wanting more.
Key Facts
- FTL is a userspace OS design
- It is compatible with Linux binaries
- The Rust-based HTTP server serving the company’s website runs on FTL
- FTL combines microkernel flexibility with monolithic kernel simplicity
- The company claims isolation is better than existing monolithic kernels
The Verdict
The design is elegant. The company’s enthusiasm is infectious. But the company has not yet shown how FTL performs at scale, how it compares to established systems, or how it handles real-world workloads.
That is a fair warning. The company is promising a new way of building cloud software, and the early signs are encouraging. The next step is to see whether the prototype survives contact with actual production traffic.
If FTL delivers on its promise, it could change how containers are treated in the cloud. If it falls short, it will join the long list of clever ideas that never quite made it out of the lab. Either way, the company has done something remarkable: they have made the operating system feel like something you can actually work with, rather than something you have to worship from afar.
Source material: “FTL: A new operating system for clouds,” ftl-os.org.
Get the Notebook.
The day's best stories and every fresh verdict, in plain English, in your inbox by seven. One email a day, no more.