A Raspberry-Pi Operating System for use in Distributed Systems

Distributed systems often come in the form of large-scale systems with expensive and proprietary hardware. These systems not only take up a large physical volume of space, but they are extremely power hungry and are typically designed with a single purpose in mind. A simple alternative to an industrial scale distributed system is a local network of computers connected via Ethernet networking. This type of setup can be used in a similar, though slower and less efficient, manner to the same ultra-scale systems. The project presented in this thesis attempts to create the base for an ultra low profile yet largely scalable version of a networked distributed system based on the modern Raspberry-Pi 4. The RPI-4 provides modern features such as high capacity memory, 64-bit architecture, gigabit networking, and massively expandable peripherals. The board features many of the advantages of embedded systems, which primarily includes physical size and versatility, while maintaining all the other modern features. The thesis lays out the development of several core features that a distributed operating system must contain. This includes advanced memory allocation, external storage access, performance measurement, and diagnostics through the use of a display monitor. These systems are integrated in a modular way with interfaces defining the interaction between them. Macros are used in places to abstract certain structures away from the programmer.