This week we learned a lot about how computers manage and store all that data they can generate. CPU and RAM are great, but once the power goes out without any persistent storage computers become a whole lot less efficient. And, well it's not just after the power goes out, but what about when a program crashes or terminates, does it need to compute everything from the start again?
In this week's course, we started from the bottom up. We learned about different I/O devices a computer can have, and how the CPU or DMA controller can communicate and save data to them. We covered a particular I/O device, a hard disk, in detail. Though hard disks with spinning disks are becoming less common today, they are still widely used in cases where capacity and longer term storage are import. We leaned a great detail about how a hard disk works physically with the head writing data to the platter and why random vs sequential has such a huge gap (due to drive head having to seek). We covered how you can use multiple disks together in RAID for different advantages. For example, RAID0 giving you a performance boost as the data is striped across multiple disks, or RAID5 or RAID6 giving you a bit of a performance boost (not quite as much as RAID0 using the same amount of drives), but possibly more important (depending on the use case) it gives you redundancy. RAID5 or RAID6 allow you to have 1 or 2 hard disks fail, respectively, and not loose any data. We then covered filesystems and how an operating system stores data to an I/O device. File systems help to organize the data and metadata in a fashion that's easy to represent to an end user.
0 comments:
Post a Comment