Mass storage system for bare machine computing

Author/Creator ORCID




Towson University. Department of Computer and Information Sciences


Citation of Original Publication




This dissertation extends on-going Bare Machine Computing (BMC) research at Towson University. BMC applications run on a bare machine without dependencies on operating systems (OS), lean kernel, embedded OS, or any other centralized support. This dissertation will serve as a cornerstone for future mass storage systems that run on bare machines . Mass storage systems based on USB flash devices are supported on many platforms and commonly used today by variety of applications. These systems vary depending upon the underlying operating system and their environments. This work leverages on the cheap flash drives with larger storage capacities. They are quite friendly to BMC as they are removable devices. Bare machine mass storage system enables computer applications to access them at run-time without the need for any operating system. This dissertation provides two avenues for mass storage systems. First, it provides a simple BMC tailored file system API for BMC programmer to store conventional file based on FAT32 specification. Second, it provides a sample database SQLite files to be stored on USBs that demonstrates the database capabilities in BMC. Both of these avenues clearly demonstrate the greater potential of BMC mass storage systems for computer applications . This dissertation developed a novel concept where a database designed on a conventional database management system is inter-operable with the database on a bare PC system. This concept is validated with SQLite database and tested for its interoperability. The idea is applicable to any other database management system as long as that system is transformed to run on a bare PC. The SQLite database management system was transformed to run on a bare PC in the earlier work as part of BMC research team. This work provides a motivation to use bare machine systems as a backend and OS based systems as a front-end to the users. That is, the actual mass storage is hidden from the users to shield from OS vulnerabilities. The lean file system interfaces for conventional files and file interface to SQLite database provides a robust and scalable mass storage systems on bare machines. Bare PCs or machines can be used for big data and clouds based on this mass storage system.