Considerations for Running the Core Flight System on Process Based Operating Systems

"NASA’s core Flight System (cFS) flight software framework was originally developed to run on real time operating systems (RTOSs) such as VxWorks in kernel mode and the Real Time Executive for Multiprocessor Systems (RTEMS). When running on these legacy real time operating systems, the cFS typically has access to the entire memory and Input/Output (I/O) address space, and applications such as Memory Dwell (MD), Checksum (CS), and Memory Manager (MM) rely on such access.

The cFS can also run on Linux using the Portable Operating System Interface (POSIX) port of the Operating System Abstraction Layer (OSAL). On Linux, the cFS runs in the context of a memory protected user process where access to system resources is restricted. The Linux port was originally intended for development and continuous integration, but it is increasingly used for cFS deployments on instruments and small satellites. In addition to Linux, there are a growing number of real time operating systems where the cFS POSIX port is applicable including QNX and the VxWorks 7 Real Time Process model. As flight systems become more complex and on-board processing capabilities grow, it will become more important for flight software to utilize these memory protected operating systems, not only for reliability, but security.

When running the cFS in the context of a memory protected process, there are considerations on how the suite of existing cFS applications will work and how the cFS can interact with the host system as a whole.

This presentation will discuss considerations for running the cFS and current suite of reusable applications in a memory protected process, and provide potential strategies for managing memory and hardware access. Additional use cases will also be explored such as running multiple instances of the cFS in separate processes on a single operating system, and modifying the cFS to support a protected application model."