CST 334 Week 7 Report

 This week, we learned a ton about the fundamentals of persistence in operating systems. We started by looking at basic device interactions with the operating system. The device itself requires two parts to make interaction efficient: the hardware interface and its internal structure. There are three main mechanisms employed to reduce CPU overhead: interruption system, programmed I/O, and direct memory access (DMA). For systems with larger memory volume transactions, DMA is superior, especially if data transactions are frequent, because the CPU does not have to be constantly used during transfers. When it comes to how the OS interacts with the device, there are two primary methods: one is to have explicit I/O instructions, and the second is known as memory-mapped I/O. Finally, the device driver is what specifically abstracts the device function away from the OS by way of a piece of software that details how a device works. Afterwards, we learned about the basics of hard disk drives. Modern disks have a platter and a spindle. Data is encoded on each surface in concentric circles of sectors that are called tracks. We read data from the surface with a disk head and arm. There are numerous disk I/O scheduling algorithms that can be employed, from more basic ones like first come, first serve to modernized algorithms like budget fair queuing. When it comes to file systems, we learned about persistent storage and devices like HDDs and SSDs. The two basic abstractions developed regarding storage are files and directories, which comprise the bread and butter of persistence in the OS. We explored the file system interface including creating, accessing, and deleting files. We wrapped up our week by learning a simple file system implementation through the vsfs (very simple file system), which is a simplified version of a typical UNIX file system. When thinking of file systems, we should be thinking about two primary aspects: the data structures of the file system and the access methods required to actually do things with data. We learned about the Inode, or index node which is the structure that holds metadata for a given file. We were able to learn more about multi-level indexing, directory organization, and free space management within a file system. All in all, it was a good week of learning and I hope we can continue to build on these basic concepts.